Research Publications
Browse peer-reviewed journal articles, conference papers, and scholarly publications. Use the filters below to refine results by type, author, year, or research area.
1.
Gita, P. C.; Arthi, R.; Geetha, R.; Hayath, S.
Exploring the Emotional Impact of Layoffs: A Twitter-Based Sentiment Analysis with NLP Techniques Proceedings
Springer Science and Business Media Deutschland GmbH, vol. 1463 LNNS, 2026, ISBN: 978-981-96-7514-2, (0).
@proceedings{712,
title = {Exploring the Emotional Impact of Layoffs: A Twitter-Based Sentiment Analysis with NLP Techniques},
author = {P. C. Gita and R. Arthi and R. Geetha and S. Hayath},
url = {https://link.springer.com/chapter/10.1007/978-981-96-7514-2_33},
doi = {10.1007/978-981-96-7514-2_33},
isbn = {978-981-96-7514-2},
year = {2026},
date = {2026-01-01},
journal = {Lecture Notes in Networks and Systems},
volume = {1463 LNNS},
pages = {413-421,},
publisher = {Springer Science and Business Media Deutschland GmbH},
abstract = {Employee layoffs have significant emotional and psychological impacts, often reflected in public discussions on social media platforms like Twitter. Employee layoffs not only affect the employee at stake but also impacts the brand image of the company. This study explores the sentiments and emotions surrounding layoffs through a comparative analysis using three natural language processing (NLP) tools: TextBlob, VADER, and the NRC Emotion Lexicon. A dataset of layoff-related tweets was collected over six months, pre-processed, and analyzed for sentiment polarity and emotional tone. The analysis revealed predominantly negative sentiments, with emotions like anger, sadness, and anticipation being prevalent. While TextBlob and VADER effectively gauged sentiment, VADER performed better in handling informal language, and the NRC Lexicon provided a more nuanced emotional profile. The study highlights the psychological toll of layoffs and the importance of employer transparency in mitigating anxiety. Future research should consider advanced NLP models like BERT for improved sentiment detection and track the evolution of layoff-related sentiments over time.},
note = {0},
keywords = {AIML},
pubstate = {published},
tppubtype = {proceedings}
}
Employee layoffs have significant emotional and psychological impacts, often reflected in public discussions on social media platforms like Twitter. Employee layoffs not only affect the employee at stake but also impacts the brand image of the company. This study explores the sentiments and emotions surrounding layoffs through a comparative analysis using three natural language processing (NLP) tools: TextBlob, VADER, and the NRC Emotion Lexicon. A dataset of layoff-related tweets was collected over six months, pre-processed, and analyzed for sentiment polarity and emotional tone. The analysis revealed predominantly negative sentiments, with emotions like anger, sadness, and anticipation being prevalent. While TextBlob and VADER effectively gauged sentiment, VADER performed better in handling informal language, and the NRC Lexicon provided a more nuanced emotional profile. The study highlights the psychological toll of layoffs and the importance of employer transparency in mitigating anxiety. Future research should consider advanced NLP models like BERT for improved sentiment detection and track the evolution of layoff-related sentiments over time.
2.
Varalatchoumy, M.; Hayath, S.; Dinesh, D.; Dhanush, C. P.; Manu, R.; Sadhana, V.
Generative AI-Powered Tool for Automated Video Summarization Proceedings
Springer Science and Business Media Deutschland GmbH, vol. 1460 LNNS, 2026, ISBN: 978-981-96-7502-9, (0).
@proceedings{686,
title = {Generative AI-Powered Tool for Automated Video Summarization},
author = {M. Varalatchoumy and S. Hayath and D. Dinesh and C. P. Dhanush and R. Manu and V. Sadhana},
url = {https://link.springer.com/chapter/10.1007/978-981-96-7502-9_10},
doi = {10.1007/978-981-96-7502-9_10},
isbn = {978-981-96-7502-9},
year = {2026},
date = {2026-01-01},
journal = {Lecture Notes in Networks and Systems},
volume = {1460 LNNS},
pages = {121-133,},
publisher = {Springer Science and Business Media Deutschland GmbH},
abstract = {This paper presents an advanced Generative AI-powered system for video-to-text summarization, leveraging state-of-the-art Computer Vision (CV) technologies and Natural Language Processing (NLP) techniques. The developed system addresses the growing need to extract key information efficiently from lengthy videos across diverse domains such as education, entertainment, sports, and instructional content. By integrating visual and textual data, it pinpoints essential moments and generates concise summaries that capture the core message of the video, reducing the time users spend understanding extensive media. At the heart of this system lies a robust, open-source large language model (LLM), fine-tuned to produce human-like summaries from video transcripts. The system processes visual cues using advanced CV techniques—such as keyframe extraction and scene segmentation—and textual cues via Automatic Speech Recognition (ASR), which converts audio into text. This dual approach facilitates a deep understanding of spoken and visual content, ensuring that summaries are precise, relevant, and contextually accurate. The system has been evaluated on a diverse dataset, comprising videos of various genres, qualities, and lengths, demonstrating its capability to generalize effectively across a wide spectrum of content. Applications of this video summarization tool include content management, video indexing, educational platforms, and beyond, offering significant time-saving benefits to users and organizations. By incorporating real-time feedback, the system continuously refines its summarization techniques, enhancing accuracy and ensuring that users quickly access the most relevant information, thereby promoting greater accessibility and usability of video content.},
note = {0},
keywords = {AIML},
pubstate = {published},
tppubtype = {proceedings}
}
This paper presents an advanced Generative AI-powered system for video-to-text summarization, leveraging state-of-the-art Computer Vision (CV) technologies and Natural Language Processing (NLP) techniques. The developed system addresses the growing need to extract key information efficiently from lengthy videos across diverse domains such as education, entertainment, sports, and instructional content. By integrating visual and textual data, it pinpoints essential moments and generates concise summaries that capture the core message of the video, reducing the time users spend understanding extensive media. At the heart of this system lies a robust, open-source large language model (LLM), fine-tuned to produce human-like summaries from video transcripts. The system processes visual cues using advanced CV techniques—such as keyframe extraction and scene segmentation—and textual cues via Automatic Speech Recognition (ASR), which converts audio into text. This dual approach facilitates a deep understanding of spoken and visual content, ensuring that summaries are precise, relevant, and contextually accurate. The system has been evaluated on a diverse dataset, comprising videos of various genres, qualities, and lengths, demonstrating its capability to generalize effectively across a wide spectrum of content. Applications of this video summarization tool include content management, video indexing, educational platforms, and beyond, offering significant time-saving benefits to users and organizations. By incorporating real-time feedback, the system continuously refines its summarization techniques, enhancing accuracy and ensuring that users quickly access the most relevant information, thereby promoting greater accessibility and usability of video content.
3.
M.S., Guru; H.N., Naveen; Jain, Amith; Syed, Javed; Baig, Rahmath
In: Journal of Food Composition and Analysis, vol. 143, pp. 107577+, 2025, ISBN: 0889-1575.
@article{392,
title = {Convergence of improved particle swarm optimization based ensemble model and explainable AI for the accurate detection of food adulteration in red chilli powder},
author = {Guru M.S. and Naveen H.N. and Amith Jain and Javed Syed and Rahmath Baig},
url = {https://www.sciencedirect.com/science/article/pii/S0889157525003928},
doi = {10.1016/j.jfca.2025.107577},
isbn = {0889-1575},
year = {2025},
date = {2025-07-01},
journal = {Journal of Food Composition and Analysis},
volume = {143},
pages = {107577+},
publisher = {Elsevier},
abstract = {Food adulteration involves the practice of adding or mixing inferior substances to food products, which undermines quality and safety. Adulteration of red chilli with brick powder is a significant food safety issue as it poses serious health risks to consumers. Accurate identification of the adulteration presents a significant challenge, particularly when adulteration is present in minuscule amounts. Existing methods aimed at identifying such micro levels of food adulteration are less accurate and lack interpretability. This study aims to address the research gaps in food adulteration by developing a robust model that integrates machine learning and explainable artificial intelligence methods. The key contributions of the proposed work are a deep convolutional generative adversarial network to enhance the model performance in limited data scenarios; improved particle swarm optimization as a promising metaheuristic optimization method to select the robust and highly discriminative features and to address premature convergence; explainable artificial intelligence methods (SHAP and LIME) to enhance the ensemble stacking model transparency and interpretability. A custom dataset is generated in the work, and it comprises 250 natural samples distributed among 5 categories (50 samples per category), ranging from no adulteration to adulteration in varying concentrations of 1 %, 2 %, 3 %, and 4 %, respectively. The proposed work is implemented on the synthetic data (200 samples per category) generated by the deep convolutional generative adversarial network. The potent combination of improved particle swarm optimization and explainable artificial intelligence enhances the accuracy, interpretability, and transparency of the proposed model by providing deeper insights which in turn bolsters confidence in distinguishing between pure and adulterated red chilli powder samples, thus contributing to improved food safety measures. The proposed model has shown a remarkable accuracy of 92.42 % on the synthetic data.},
keywords = {AIML},
pubstate = {published},
tppubtype = {article}
}
Food adulteration involves the practice of adding or mixing inferior substances to food products, which undermines quality and safety. Adulteration of red chilli with brick powder is a significant food safety issue as it poses serious health risks to consumers. Accurate identification of the adulteration presents a significant challenge, particularly when adulteration is present in minuscule amounts. Existing methods aimed at identifying such micro levels of food adulteration are less accurate and lack interpretability. This study aims to address the research gaps in food adulteration by developing a robust model that integrates machine learning and explainable artificial intelligence methods. The key contributions of the proposed work are a deep convolutional generative adversarial network to enhance the model performance in limited data scenarios; improved particle swarm optimization as a promising metaheuristic optimization method to select the robust and highly discriminative features and to address premature convergence; explainable artificial intelligence methods (SHAP and LIME) to enhance the ensemble stacking model transparency and interpretability. A custom dataset is generated in the work, and it comprises 250 natural samples distributed among 5 categories (50 samples per category), ranging from no adulteration to adulteration in varying concentrations of 1 %, 2 %, 3 %, and 4 %, respectively. The proposed work is implemented on the synthetic data (200 samples per category) generated by the deep convolutional generative adversarial network. The potent combination of improved particle swarm optimization and explainable artificial intelligence enhances the accuracy, interpretability, and transparency of the proposed model by providing deeper insights which in turn bolsters confidence in distinguishing between pure and adulterated red chilli powder samples, thus contributing to improved food safety measures. The proposed model has shown a remarkable accuracy of 92.42 % on the synthetic data.
4.
Minh, N. V. Van; Poorani, B.; Vijayakumar, R.; Varalatchoumy, M.; Sreevidya, R. C.; Prabhakar, P.
In: Thermal Science and Engineering Progress, vol. 68, 2025, ISBN: 24519049 (ISSN), (0).
@article{689,
title = {Experimental and machine learning-based analysis of peanut drying using solar Photovoltaic-Thermal (PVT) collectors with forced convection and latent heat storage},
author = {N. V. Van Minh and B. Poorani and R. Vijayakumar and M. Varalatchoumy and R. C. Sreevidya and P. Prabhakar},
url = {https://www.sciencedirect.com/science/article/pii/S2451904925011254?via%3Dihub},
doi = {10.1016/j.tsep.2025.104334},
isbn = {24519049 (ISSN)},
year = {2025},
date = {2025-01-01},
journal = {Thermal Science and Engineering Progress},
volume = {68},
publisher = {Elsevier Ltd},
abstract = {This paper provides a detailed experimental and machine learning analysis of peanut drying with a hybrid solar photovoltaic-thermal (PVT) collector system. Four drying techniques were tested: open sun drying, natural convection, forced convection, and forced convection combined with a paraffin wax-based Phase Change Material (PCM) to store latent heat. Each condition involved drying 5 kg of wet peanuts with an initial moisture content of 40 %. The drying processes were simulated using three machine learning models Gaussian Process Regression (GPR), Radial Basis Function (RBF), and Multilayer Perceptron (MLP) to predict moisture removal and drying performance. Model accuracy was measured by RMSE, MAPE, and R2. The results show that forced convection with PCM was the most successful approach, lowering drying time from around 42 h (open sun drying) to only 18 h and attaining the best drying efficiency of 68.23 %. The greatest electrical efficiency was 11.24 %, while the collector efficiency was 21.89 %. The RBF network outperformed the GPR and MLP models in moisture removal and drying performance, with R2 values of 0.97 and 0.98, respectively. This study concludes that integrating PCM with forced convection in a PVT-dryer system, together with powerful machine learning predictions, provides a highly efficient and sustainable approach for agricultural product preservation that outperforms existing techniques.},
note = {0},
keywords = {AIML},
pubstate = {published},
tppubtype = {article}
}
This paper provides a detailed experimental and machine learning analysis of peanut drying with a hybrid solar photovoltaic-thermal (PVT) collector system. Four drying techniques were tested: open sun drying, natural convection, forced convection, and forced convection combined with a paraffin wax-based Phase Change Material (PCM) to store latent heat. Each condition involved drying 5 kg of wet peanuts with an initial moisture content of 40 %. The drying processes were simulated using three machine learning models Gaussian Process Regression (GPR), Radial Basis Function (RBF), and Multilayer Perceptron (MLP) to predict moisture removal and drying performance. Model accuracy was measured by RMSE, MAPE, and R2. The results show that forced convection with PCM was the most successful approach, lowering drying time from around 42 h (open sun drying) to only 18 h and attaining the best drying efficiency of 68.23 %. The greatest electrical efficiency was 11.24 %, while the collector efficiency was 21.89 %. The RBF network outperformed the GPR and MLP models in moisture removal and drying performance, with R2 values of 0.97 and 0.98, respectively. This study concludes that integrating PCM with forced convection in a PVT-dryer system, together with powerful machine learning predictions, provides a highly efficient and sustainable approach for agricultural product preservation that outperforms existing techniques.
5.
Poorani, B.; Sreevidya, R. C.; Vijayakumar, R.; Varalatchoumy, M.; Natesan, P.; Prabhakar, P.
Experimental exergy analysis of SnO2 nanofluid photovoltaic thermal system using machine learning approach Journal Article
In: Journal of Thermal Analysis and Calorimetry, 2025, ISBN: 13886150 (ISSN); 15882926 (ISSN), (0).
@article{697,
title = {Experimental exergy analysis of SnO2 nanofluid photovoltaic thermal system using machine learning approach},
author = {B. Poorani and R. C. Sreevidya and R. Vijayakumar and M. Varalatchoumy and P. Natesan and P. Prabhakar},
url = {https://link.springer.com/article/10.1007/s10973-025-14938-7},
doi = {10.1007/s10973-025-14938-7},
isbn = {13886150 (ISSN); 15882926 (ISSN)},
year = {2025},
date = {2025-01-01},
journal = {Journal of Thermal Analysis and Calorimetry},
publisher = {Springer Science and Business Media B.V.},
abstract = {The efficiency of photovoltaic thermal (PVT) systems is often hindered by high operating temperatures, which can be effectively addressed through advanced cooling methods. This study explored the use of a water-based tin dioxide (SnO<inf>2</inf>) nanofluid at a 0.1% concentration as an enhanced coolant to boost the system’s exergy efficiency. The research involved experimental testing under three distinct flow rates—0.5, 1.0 and 1.5 LPM—to evaluate the nanofluid’s performance. The results confirmed that the nanofluid offered a significant advantage over conventional pure water cooling. Specifically, at the highest flow rate of 1.5 LPM, the maximum exergy efficiency improved remarkably from 11.1 to 18.9%. In addition to the experimental work, the study also developed and tested several machine learning (ML) models to predict the system’s performance. Two primary models, K-Nearest Neighbor (KNN) and Support Vector Regression (SVR), were utilized. The researchers also investigated the impact of integrating Wavelet Transform (WT), a signal-processing technique, with these ML models. The results demonstrated that the SVR model combined with Wavelet Transform (SVR-WT) provided the most accurate predictions on the test dataset. This model achieved an impressive coefficient of determination (R2) of 0.885, indicating a strong correlation between the predicted and actual values. Its predictive capability was further highlighted by a low root mean square error (RMSE) of 2.196 and a mean absolute error (MAE) of 3.086. Overall, the findings conclusively establish that SnO2 nanofluid is an excellent coolant for enhancing PVT system performance, and that the SVR-WT model offers a reliable predictive framework for optimizing these systems.},
note = {0},
keywords = {AIML},
pubstate = {published},
tppubtype = {article}
}
The efficiency of photovoltaic thermal (PVT) systems is often hindered by high operating temperatures, which can be effectively addressed through advanced cooling methods. This study explored the use of a water-based tin dioxide (SnO<inf>2</inf>) nanofluid at a 0.1% concentration as an enhanced coolant to boost the system’s exergy efficiency. The research involved experimental testing under three distinct flow rates—0.5, 1.0 and 1.5 LPM—to evaluate the nanofluid’s performance. The results confirmed that the nanofluid offered a significant advantage over conventional pure water cooling. Specifically, at the highest flow rate of 1.5 LPM, the maximum exergy efficiency improved remarkably from 11.1 to 18.9%. In addition to the experimental work, the study also developed and tested several machine learning (ML) models to predict the system’s performance. Two primary models, K-Nearest Neighbor (KNN) and Support Vector Regression (SVR), were utilized. The researchers also investigated the impact of integrating Wavelet Transform (WT), a signal-processing technique, with these ML models. The results demonstrated that the SVR model combined with Wavelet Transform (SVR-WT) provided the most accurate predictions on the test dataset. This model achieved an impressive coefficient of determination (R2) of 0.885, indicating a strong correlation between the predicted and actual values. Its predictive capability was further highlighted by a low root mean square error (RMSE) of 2.196 and a mean absolute error (MAE) of 3.086. Overall, the findings conclusively establish that SnO2 nanofluid is an excellent coolant for enhancing PVT system performance, and that the SVR-WT model offers a reliable predictive framework for optimizing these systems.
6.
Krishnamoorthy, N.; Varalatchoumy, M.; Aruna, R.; Satyanarayana, G.; Karthikeyan, P.; Kumar, E.
High-Efficiency Triple-Band Antenna Design for Next-Generation Wireless Technologies Journal Article
In: Internet Technology Letters, vol. 8, 2025, ISBN: 24761508 (ISSN), (0).
@article{632,
title = {High-Efficiency Triple-Band Antenna Design for Next-Generation Wireless Technologies},
author = {N. Krishnamoorthy and M. Varalatchoumy and R. Aruna and G. Satyanarayana and P. Karthikeyan and E. Kumar},
url = {https://onlinelibrary.wiley.com/doi/10.1002/itl2.70078},
doi = {10.1002/itl2.70078},
isbn = {24761508 (ISSN)},
year = {2025},
date = {2025-01-01},
journal = {Internet Technology Letters},
volume = {8},
publisher = {John Wiley and Sons Inc},
abstract = {This paper presents a compact antenna design with a tailored ground structure optimized for triple-band wireless applications. The proposed antenna operates efficiently across three distinct frequency bands: 1.9–5 GHz, 6.1–7.4 GHz, and 8.5–10.2 GHz, making it ideal for emerging wireless technologies including 5G, Wi-Fi 6E, and X-band communications. Comprehensive design optimization yields consistent S<inf>11</inf> values below −10 dB within these bands, delivering bandwidths of 3.1, 1.3, and 1.7 GHz, respectively. The antenna achieves a peak radiation efficiency of 80% and a gain of 4.5 dBi, all within a simple and compact structure suitable for versatile wireless applications.},
note = {0},
keywords = {AIML},
pubstate = {published},
tppubtype = {article}
}
This paper presents a compact antenna design with a tailored ground structure optimized for triple-band wireless applications. The proposed antenna operates efficiently across three distinct frequency bands: 1.9–5 GHz, 6.1–7.4 GHz, and 8.5–10.2 GHz, making it ideal for emerging wireless technologies including 5G, Wi-Fi 6E, and X-band communications. Comprehensive design optimization yields consistent S<inf>11</inf> values below −10 dB within these bands, delivering bandwidths of 3.1, 1.3, and 1.7 GHz, respectively. The antenna achieves a peak radiation efficiency of 80% and a gain of 4.5 dBi, all within a simple and compact structure suitable for versatile wireless applications.
7.
Ramkumar, M. Siva; Senthilvel, N.; Arun, Mvarun; Baig, Rahmath; Giri, Jayant; Al-Qawasmi, Khaled
2025.
@book{596,
title = {Classification and Optimization of Liver Cancer Detection using Capsule-Vectored Neural Network with Artificial Jelly Optimization},
author = {M. Siva Ramkumar and N. Senthilvel and Mvarun Arun and Rahmath Baig and Jayant Giri and Khaled Al-Qawasmi},
url = {https://ieeexplore.ieee.org/document/11070423},
doi = {10.1109/ICDSIS65355.2025.11070423},
year = {2025},
date = {2025-01-01},
pages = {1-6,},
abstract = {Liver cancer diagnostic procedures currently use CT together with MRI and ultrasonic scans to check for tumors while several diagnostic challenges like tissue superimposition and irregular tumor formation limit test precision. The proposed solution involves deploying CV2N-AJOpt which stands for Capsule-Vectored Neural Network with Artificial Jelly Optimization. The LiTS dataset receives preprocess treatment using Gradient Domain Guided Filtering (2GDF) to improve image quality and remove noise. STFT-CV2Net serves as the combination of Short-Time Fourier Transform and Capsule-Vectored Neural Network to execute feature extraction and classification operations. AJO performs weight optimization of CV2Net to enhance its operational performance. The proposed model delivers experimental results showing 99.8% recall and 99.9% accuracy which establishes superiority compared to existing methods. CV2N-AJOpt shows excellence at reducing wrong detections while developing high accuracy in detecting complex liver cancer anatomical structures. STFT-CV2Net demonstrates value as a clinical diagnosis tool for liver cancer because it effectively processes extensive medical data while performing with reliability.},
keywords = {AIML},
pubstate = {published},
tppubtype = {book}
}
Liver cancer diagnostic procedures currently use CT together with MRI and ultrasonic scans to check for tumors while several diagnostic challenges like tissue superimposition and irregular tumor formation limit test precision. The proposed solution involves deploying CV2N-AJOpt which stands for Capsule-Vectored Neural Network with Artificial Jelly Optimization. The LiTS dataset receives preprocess treatment using Gradient Domain Guided Filtering (2GDF) to improve image quality and remove noise. STFT-CV2Net serves as the combination of Short-Time Fourier Transform and Capsule-Vectored Neural Network to execute feature extraction and classification operations. AJO performs weight optimization of CV2Net to enhance its operational performance. The proposed model delivers experimental results showing 99.8% recall and 99.9% accuracy which establishes superiority compared to existing methods. CV2N-AJOpt shows excellence at reducing wrong detections while developing high accuracy in detecting complex liver cancer anatomical structures. STFT-CV2Net demonstrates value as a clinical diagnosis tool for liver cancer because it effectively processes extensive medical data while performing with reliability.
8.
Al-Shaikh, A.; H, G.; Singh, N.; .Sivaramkrishnan, M; .Dhivya, S; Ramkumar, M.
Graph Attention Fusion Network for Electric Vehicle Charging Stations Management in Vehicle-to-Grid Systems Proceedings
2025.
@proceedings{595,
title = {Graph Attention Fusion Network for Electric Vehicle Charging Stations Management in Vehicle-to-Grid Systems},
author = {A. Al-Shaikh and G. H and N. Singh and M .Sivaramkrishnan and S .Dhivya and M. Ramkumar},
url = {https://ieeexplore.ieee.org/document/11074140},
doi = {10.1109/ICIMA64861.2025.11074140},
year = {2025},
date = {2025-01-01},
journal = {2025 7th International Conference on Inventive Material Science and Applications (ICIMA)},
pages = {586-592,},
abstract = {Vehicle-to-Grid (V2G) systems enable bidirectional energy flow between Electric Vehicles (EVs) and the grid, facilitating Energy Management (EM) and grid stability. Effective management of EV charging stations within V2G networks optimizes power distribution and supports renewable energy integration. However, high infrastructure costs associated with bidirectional chargers and grid upgrades pose financial challenges. Additionally, frequent charging and discharging cycles can lead to increased battery degradation, indirectly contributing to higher emissions from battery production and disposal. To overcome these drawbacks, this manuscript proposes a method for EV Charging Stations (EVCSs) in V2G systems. The proposed method is Graph Attention Fusion Network (GAF-Net). The main aim of the proposed method is to reduce the operational cost, charging time and emission of the system. The proposed GAF-Net predicts energy demand patterns and renewable energy generation, facilitating proactive decision-making for efficient management of EVCSs. Artificial Neural Network (ANN), Assailant Inspired Chimp Optimization Algorithm (AIChOa), and Non-Dominated Sorting Genetic Algorithm-II (NSGA-II) are some of the existing techniques that are compared with the proposed method once it is implemented in MATLAB. By achieving the lowest cost of 1504cents and the lowest emission of 60.4ppm, the proposed GAF-Net technique outperforms existing methods while ensuring cost-effective and efficient EVCS management in V2G systems integrating Photovoltaic (PV) systems and battery storage.},
keywords = {AIML},
pubstate = {published},
tppubtype = {proceedings}
}
Vehicle-to-Grid (V2G) systems enable bidirectional energy flow between Electric Vehicles (EVs) and the grid, facilitating Energy Management (EM) and grid stability. Effective management of EV charging stations within V2G networks optimizes power distribution and supports renewable energy integration. However, high infrastructure costs associated with bidirectional chargers and grid upgrades pose financial challenges. Additionally, frequent charging and discharging cycles can lead to increased battery degradation, indirectly contributing to higher emissions from battery production and disposal. To overcome these drawbacks, this manuscript proposes a method for EV Charging Stations (EVCSs) in V2G systems. The proposed method is Graph Attention Fusion Network (GAF-Net). The main aim of the proposed method is to reduce the operational cost, charging time and emission of the system. The proposed GAF-Net predicts energy demand patterns and renewable energy generation, facilitating proactive decision-making for efficient management of EVCSs. Artificial Neural Network (ANN), Assailant Inspired Chimp Optimization Algorithm (AIChOa), and Non-Dominated Sorting Genetic Algorithm-II (NSGA-II) are some of the existing techniques that are compared with the proposed method once it is implemented in MATLAB. By achieving the lowest cost of 1504cents and the lowest emission of 60.4ppm, the proposed GAF-Net technique outperforms existing methods while ensuring cost-effective and efficient EVCS management in V2G systems integrating Photovoltaic (PV) systems and battery storage.
9.
Priyanka, R.; Karaguppi, M. R.; Shree, Ramya; Shankar, Gowri
Medibot: An AI-Enabled IoT-Based Smart Pill Dispenser with App-Based Dosage and Expiration Alerts Proceedings
Institute of Electrical and Electronics Engineers Inc., 2025, ISBN: 979-833150767-1 (ISBN), (0).
@proceedings{448,
title = {Medibot: An AI-Enabled IoT-Based Smart Pill Dispenser with App-Based Dosage and Expiration Alerts},
author = {R. Priyanka and M. R. Karaguppi and Ramya Shree and Gowri Shankar},
url = {https://ieeexplore.ieee.org/document/11009992},
doi = {10.1109/ICSSES64899.2025.11009992},
isbn = {979-833150767-1 (ISBN)},
year = {2025},
date = {2025-01-01},
journal = {International Conference on Smart Systems for Applications in Electrical Sciences, ICSSES 2025},
publisher = {Institute of Electrical and Electronics Engineers Inc.},
abstract = {An innovative healthcare solution, the 'Medibot' is made to make medication management easier and more efficient, especially for elderly and ill patients. It uses voice recognition and location-based movement to deliver the right dosage of the drug and water at the right time, making it more accessible to the patient. It is integrated through a mobile application to facilitate easy scheduling, monitoring, and alerting mechanisms; it has biometric security, low-stock notification, and expiration date scanning to ensure safety and medication adherence; it has a scanner and navigation feature for improved accessibility. This sophisticated technology seeks to enhance patient health, lower the chance of missing or expired medicine, and make it easier for carers to provide effective patient care. Better drug regimens are exemplified by the integration of several hardware and software technologies, which results in a dependable, safe, and easily accessible healthcare solution.},
note = {0},
keywords = {AIML},
pubstate = {published},
tppubtype = {proceedings}
}
An innovative healthcare solution, the 'Medibot' is made to make medication management easier and more efficient, especially for elderly and ill patients. It uses voice recognition and location-based movement to deliver the right dosage of the drug and water at the right time, making it more accessible to the patient. It is integrated through a mobile application to facilitate easy scheduling, monitoring, and alerting mechanisms; it has biometric security, low-stock notification, and expiration date scanning to ensure safety and medication adherence; it has a scanner and navigation feature for improved accessibility. This sophisticated technology seeks to enhance patient health, lower the chance of missing or expired medicine, and make it easier for carers to provide effective patient care. Better drug regimens are exemplified by the integration of several hardware and software technologies, which results in a dependable, safe, and easily accessible healthcare solution.
10.
Ramkumar, M.; Sivaramkrishnan, M.; M, A.; S, Y.; Jawad, O.; Chadge, R.
2025.
@proceedings{594,
title = {Enhanced EV Spatial Distribution Estimation for Grid Planning using Neural Architecture Search-Guided Physics-Informed Neural Network and Pufferfish Optimization Algorithm},
author = {M. Ramkumar and M. Sivaramkrishnan and A. M and Y. S and O. Jawad and R. Chadge},
url = {https://ieeexplore.ieee.org/document/11074243},
doi = {10.1109/ICIMA64861.2025.11074243},
year = {2025},
date = {2025-01-01},
journal = {2025 7th International Conference on Inventive Material Science and Applications (ICIMA)},
pages = {460-466,},
abstract = {Electric vehicles (EV) play a vital role in modern transportation, and accurately estimating their spatial distribution is essential for effective grid planning. The utility grid management system will optimize energy utilization while balancing network capacity and develop new infrastructure components which reduce the electrical grid stress during EV integration increases. However, estimating EV distribution in grid planning is challenging due to unpredictable charging patterns, dynamic mobility behavior, and varying energy demands, which complicate load forecasting and infrastructure optimization. To overcome these drawbacks, this paper proposes a hybrid approach for spatial distribution of EV. The process begins by gathering data from vehicle registration dataset, which is then passed through a pre-processing phase. Regularized Bias-Aware Ensemble Kalman Filter (RBAEKF) is employed to clean and remove the missing value in the input data. The pre-processed output was fed to Neural Architecture Search-Guided Physics-Informed Neural Network (NASPINN) the data enters the classification phase, to enhance the accuracy of classifications. The class 0, class 1 and class 2 of EV distribution is successfully classified by using NASPINN. The weight parameter of NASPINN is optimized using Pufferfish Optimization Algorithm (POA). The NASPINN-POA technique is implemented in MATLAB and evaluated using various performance metrics, including accuracy, precision, recall, F1-score, specificity and Root Mean Squared Error (RMSE). The results show that the NASPINN-POA method outperforms existing approaches, such as Sparrow Search Algorithm-Back Propagation Neural Network (SSA-BPNN), Particle Swarm Optimization (PSO), Eurasian Oystercatcher Optimizer-Quantum Neural Network (EOO-QNN), Bayesian Network (BN) and Long Short-Term Memory (LSTM). The proposed NASPINN-POA method enables an accurate spatial distribution of EV with 98.8% accuracy, 98.4% recall and achieves an MAE of 1.07 to optimize grid planning by minimizing errors.},
keywords = {AIML},
pubstate = {published},
tppubtype = {proceedings}
}
Electric vehicles (EV) play a vital role in modern transportation, and accurately estimating their spatial distribution is essential for effective grid planning. The utility grid management system will optimize energy utilization while balancing network capacity and develop new infrastructure components which reduce the electrical grid stress during EV integration increases. However, estimating EV distribution in grid planning is challenging due to unpredictable charging patterns, dynamic mobility behavior, and varying energy demands, which complicate load forecasting and infrastructure optimization. To overcome these drawbacks, this paper proposes a hybrid approach for spatial distribution of EV. The process begins by gathering data from vehicle registration dataset, which is then passed through a pre-processing phase. Regularized Bias-Aware Ensemble Kalman Filter (RBAEKF) is employed to clean and remove the missing value in the input data. The pre-processed output was fed to Neural Architecture Search-Guided Physics-Informed Neural Network (NASPINN) the data enters the classification phase, to enhance the accuracy of classifications. The class 0, class 1 and class 2 of EV distribution is successfully classified by using NASPINN. The weight parameter of NASPINN is optimized using Pufferfish Optimization Algorithm (POA). The NASPINN-POA technique is implemented in MATLAB and evaluated using various performance metrics, including accuracy, precision, recall, F1-score, specificity and Root Mean Squared Error (RMSE). The results show that the NASPINN-POA method outperforms existing approaches, such as Sparrow Search Algorithm-Back Propagation Neural Network (SSA-BPNN), Particle Swarm Optimization (PSO), Eurasian Oystercatcher Optimizer-Quantum Neural Network (EOO-QNN), Bayesian Network (BN) and Long Short-Term Memory (LSTM). The proposed NASPINN-POA method enables an accurate spatial distribution of EV with 98.8% accuracy, 98.4% recall and achieves an MAE of 1.07 to optimize grid planning by minimizing errors.
11.
M., Sivaram; R, Geetha; Gopalakrishnan, Emayavarmban; Ramkumar, M. Siva; Abu-Saleem, Thaer; Chadge, Rajkumar
Spider Wasp Optimization-based Power Management for Connected Electric Vehicles Book
2025.
@book{593,
title = {Spider Wasp Optimization-based Power Management for Connected Electric Vehicles},
author = {Sivaram M. and Geetha R and Emayavarmban Gopalakrishnan and M. Siva Ramkumar and Thaer Abu-Saleem and Rajkumar Chadge},
url = {https://ieeexplore.ieee.org/document/11074141},
doi = {10.1109/ICIMA64861.2025.11074141},
year = {2025},
date = {2025-01-01},
pages = {449-454,},
abstract = {Power Management (PM) in connected Electric Vehicles (EVs) entails coordinating the energy flow between the vehicles, charging stations, and the power grid in such a way as to achieve optimal functioning. Yet, energy supply in connected EVs would remain tricky, especially with the grid's constraints, variability in renewable energy generation, and variation in charging demand. This paper proposes a Spider Wasp Optimization (SWO) for enhancing PM in connected EVs. The SWO is used to minimize Cost of Energy (COE) in EVs. By optimizing the operational parameters, SWO minimizes the energy cost involved in charging EVs while maximizing the use of PV energy and managing energy distribution from various sources. Its adaptive optimization ensures that COE is minimized across various operating conditions, enhancing overall efficiency and responsiveness of PM of EVs. By then the proposed SWO method is implemented in MATLAB platform and evaluated their performance with various existing methods such as Multi Island Genetic Algorithm (MIGA), Genetic Algorithm (GA), Deep Deterministic Policy Gradient Algorithm (D3PG), Stochastic Multi Objective Optimization (SMOO), and Grey Sail Fish Optimization (GSFO). The proposed SWO method outperforms all the others with the lowest COE of $0.043/kWh, indicating its superior performance cost effectiveness of PM in connected EVs.},
keywords = {AIML},
pubstate = {published},
tppubtype = {book}
}
Power Management (PM) in connected Electric Vehicles (EVs) entails coordinating the energy flow between the vehicles, charging stations, and the power grid in such a way as to achieve optimal functioning. Yet, energy supply in connected EVs would remain tricky, especially with the grid's constraints, variability in renewable energy generation, and variation in charging demand. This paper proposes a Spider Wasp Optimization (SWO) for enhancing PM in connected EVs. The SWO is used to minimize Cost of Energy (COE) in EVs. By optimizing the operational parameters, SWO minimizes the energy cost involved in charging EVs while maximizing the use of PV energy and managing energy distribution from various sources. Its adaptive optimization ensures that COE is minimized across various operating conditions, enhancing overall efficiency and responsiveness of PM of EVs. By then the proposed SWO method is implemented in MATLAB platform and evaluated their performance with various existing methods such as Multi Island Genetic Algorithm (MIGA), Genetic Algorithm (GA), Deep Deterministic Policy Gradient Algorithm (D3PG), Stochastic Multi Objective Optimization (SMOO), and Grey Sail Fish Optimization (GSFO). The proposed SWO method outperforms all the others with the lowest COE of $0.043/kWh, indicating its superior performance cost effectiveness of PM in connected EVs.
12.
Ramkumar, M. Siva; Kavitha, D.; Arun, Mvarun; M, Varalatchoumy; Abu-Saleem, Thaer; Chadge, Rajkumar
2025.
@book{592,
title = {Ultracapacitor-Assisted Energy Management in Hybrid Electric Vehicles using QSNGNN: A Quaternion Similarity-based GNN Approach},
author = {M. Siva Ramkumar and D. Kavitha and Mvarun Arun and Varalatchoumy M and Thaer Abu-Saleem and Rajkumar Chadge},
url = {https://ieeexplore.ieee.org/document/11073951},
doi = {10.1109/ICIMA64861.2025.11073951},
year = {2025},
date = {2025-01-01},
pages = {301-307,},
abstract = {An effective Energy Management (EM) system within Hybrid Electric Vehicles (HEVs) with Fuel Cells (FCs) and batteries together with Ultracapacitors (UCs) optimizes power distribution to enhance hydrogen usage. The inefficiency of power distribution leads to enhanced hydrogen usage but leads to control system delays that negatively impact total system performance. To overcome these drawbacks, this manuscript proposes an approach for EM of HEV with FC, battery, and UC. The suggested method is Starfish Optimization Algorithm (SfOA). The main aim of the suggested method is to enhance energy efficiency, reduces hydrogen consumption, and improves the vehicle s overall performance. SFO optimizes power distribution among the FC, battery, and UCs to minimize hydrogen consumption and regulate energy flow efficiently. By then, the suggested approach is implemented in MATLAB and contrasted with several other methods that are previously in use. The suggested technique outperforms all previous techniques such as Artificial Neural Network (ANN), Convolutional Neural Network (CNN), Twin Delayed Deep Deterministic Policy Gradient Algorithm (TD3PGA), Cheetah Optimizer-Spiking Neural Network (CO-SNN), and Radial Basis Function Neural Network (RBFNN). The SfOA method shows operational effectiveness of 98.7% and consumes 22.71g of hydrogen during operation. The suggested method provides higher energy efficiency and reduced hydrogen consumption which establishes it as an efficient solution for EM in HEVs with hybrid powertrains involving FCs, batteries, and UCs.},
keywords = {AIML},
pubstate = {published},
tppubtype = {book}
}
An effective Energy Management (EM) system within Hybrid Electric Vehicles (HEVs) with Fuel Cells (FCs) and batteries together with Ultracapacitors (UCs) optimizes power distribution to enhance hydrogen usage. The inefficiency of power distribution leads to enhanced hydrogen usage but leads to control system delays that negatively impact total system performance. To overcome these drawbacks, this manuscript proposes an approach for EM of HEV with FC, battery, and UC. The suggested method is Starfish Optimization Algorithm (SfOA). The main aim of the suggested method is to enhance energy efficiency, reduces hydrogen consumption, and improves the vehicle s overall performance. SFO optimizes power distribution among the FC, battery, and UCs to minimize hydrogen consumption and regulate energy flow efficiently. By then, the suggested approach is implemented in MATLAB and contrasted with several other methods that are previously in use. The suggested technique outperforms all previous techniques such as Artificial Neural Network (ANN), Convolutional Neural Network (CNN), Twin Delayed Deep Deterministic Policy Gradient Algorithm (TD3PGA), Cheetah Optimizer-Spiking Neural Network (CO-SNN), and Radial Basis Function Neural Network (RBFNN). The SfOA method shows operational effectiveness of 98.7% and consumes 22.71g of hydrogen during operation. The suggested method provides higher energy efficiency and reduced hydrogen consumption which establishes it as an efficient solution for EM in HEVs with hybrid powertrains involving FCs, batteries, and UCs.
13.
C, Rajendran; Balassem, Zaid; S, Nandini; Hayath, Syed; V, Suma; Biswas, Debarghya
Dynamic Load Balancing in Dense Urban Mobile Networks Journal Article
In: vol. 15, pp. 447-458,, 2025.
@article{591,
title = {Dynamic Load Balancing in Dense Urban Mobile Networks},
author = {Rajendran C and Zaid Balassem and Nandini S and Syed Hayath and Suma V and Debarghya Biswas},
url = {https://jisis.org/wp-content/uploads/2025/07/2025.I2.032.pdf},
doi = {10.58346/JISIS.2025.I2.032},
year = {2025},
date = {2025-01-01},
volume = {15},
pages = {447-458,},
abstract = {The increasing number of mobile users and data-reliant applications in a densely populated area has created the need for better mobile networks. Static and semi-dynamic load distribution methods are usually incapable of coping with the pace of urban mobility and the evolving demands of a network. This paper proposes a dynamic load-balancing approach to improve resource allocation within heterogeneous mobile networks utilizing user mobility patterns, real-time traffic data, adaptive handover methods, and user-guided handover escalation policies. Selective handover initiation and load-shifting decisions are made from a continual assessment of network load vis-a-vis user distribution. Simulation has shown improved results over traditional methods concerning network performance metrics such as latency, throughput, and quality of service, as well as the overall satisfaction level of the users. The framework is instrumental in alleviating congestion and increasing services' reliability in highly populated urban environments, thus solving problems of 5G and next-generation mobile networks.},
keywords = {AIML},
pubstate = {published},
tppubtype = {article}
}
The increasing number of mobile users and data-reliant applications in a densely populated area has created the need for better mobile networks. Static and semi-dynamic load distribution methods are usually incapable of coping with the pace of urban mobility and the evolving demands of a network. This paper proposes a dynamic load-balancing approach to improve resource allocation within heterogeneous mobile networks utilizing user mobility patterns, real-time traffic data, adaptive handover methods, and user-guided handover escalation policies. Selective handover initiation and load-shifting decisions are made from a continual assessment of network load vis-a-vis user distribution. Simulation has shown improved results over traditional methods concerning network performance metrics such as latency, throughput, and quality of service, as well as the overall satisfaction level of the users. The framework is instrumental in alleviating congestion and increasing services' reliability in highly populated urban environments, thus solving problems of 5G and next-generation mobile networks.
14.
Ramkumar, M. Siva; Prabu, R.; Arun, Mvarun; Baig, Rahmath; P, Suseendhar; Chandravathi, C.
2025.
@book{590,
title = {Cost-Effective Optimization of Renewable Energy Systems with Hybrid Atomic Orbital and Binary Light Spectrum Algorithms},
author = {M. Siva Ramkumar and R. Prabu and Mvarun Arun and Rahmath Baig and Suseendhar P and C. Chandravathi},
url = {https://ieeexplore.ieee.org/document/11070733},
doi = {10.1109/ICDSIS65355.2025.11070733},
year = {2025},
date = {2025-01-01},
pages = {1-7,},
abstract = {This study investigates the operational performance of off-grid power supply systems incorporating renewable energy sources (RES) and storage technologies. Four system configurations single RES, double RES, single storage, and double storage are evaluated under two scenarios. A hybrid pumped battery backup system (HPBS) is proposed, where pumped hydro storage (PHS) addresses high energy fluctuations, and batteries manage minor imbalances. To optimize the system design, a hybrid Atomic Orbital Search Algorithm integrated with the Binary Light Spectrum Optimizer (AOSA-BLSO) is employed. The results indicate that self-discharge notably impacts energy cost (13% 50%) due to oversized RE generation compared to storage capacity. Among the configurations, solar-wind-PHS is the most economical, while solar-wind-HPBS offers superior reliability. Sensitivity analysis confirms the added flexibility and effectiveness of combining PHS and battery storage. The proposed optimization approach achieved 98% efficiency and reduced costs by 4000k, outperforming existing techniques. The study highlights HPBS as vital for cost-effective, reliable off-grid RE systems.},
keywords = {AIML},
pubstate = {published},
tppubtype = {book}
}
This study investigates the operational performance of off-grid power supply systems incorporating renewable energy sources (RES) and storage technologies. Four system configurations single RES, double RES, single storage, and double storage are evaluated under two scenarios. A hybrid pumped battery backup system (HPBS) is proposed, where pumped hydro storage (PHS) addresses high energy fluctuations, and batteries manage minor imbalances. To optimize the system design, a hybrid Atomic Orbital Search Algorithm integrated with the Binary Light Spectrum Optimizer (AOSA-BLSO) is employed. The results indicate that self-discharge notably impacts energy cost (13% 50%) due to oversized RE generation compared to storage capacity. Among the configurations, solar-wind-PHS is the most economical, while solar-wind-HPBS offers superior reliability. Sensitivity analysis confirms the added flexibility and effectiveness of combining PHS and battery storage. The proposed optimization approach achieved 98% efficiency and reduced costs by 4000k, outperforming existing techniques. The study highlights HPBS as vital for cost-effective, reliable off-grid RE systems.
15.
L, N.; Suneetha, J.; Latha, S.; B, S.; Tabassum, H.; N, S.
2025.
@proceedings{589,
title = {Farm-to-Folk: Leveraging Machine Learning for Efficient Agricultural Production, Supply Chain Optimization, and Sustainable Food Distribution},
author = {N. L and J. Suneetha and S. Latha and S. B and H. Tabassum and S. N},
url = {https://ieeexplore.ieee.org/document/11064218},
doi = {10.1109/ICECCC65144.2025.11064218},
year = {2025},
date = {2025-01-01},
journal = {2025 International Conference on Electronics, Computing, Communication and Control Technology (ICECCC)},
pages = {1-5,},
abstract = {Using a thorough set of environmental data, the Farm to Folk initiative forecasts the most appropriate crops for production by leveraging machine learning methods. This data set includes crucial agricultural indicators, including nitrogen levels, phosphorus content, temperature, humidity, pH levels of water, and other pertinent variables. Using advanced data processing techniques, feature engineering methodologies, and iterative model training procedures, the system attempts to accurately predict the optimal crops to be grown under specific environmental conditions. The efficacy of the predictive model is rigorously evaluated against real-world agricultural scenarios and historical crop yield data. By means of this iterative process, Farm-to-People seeks to provide farmers with a consistent decision-support tool that takes several environmental factors into account, thereby empowering them to make wise crop choice. By facilitating precision agriculture through machine learning-driven predictions, this initiative ultimately aims to preagricultural efficiency, optimize resource utilization, and propromoteustainable farming practices.},
keywords = {AIML},
pubstate = {published},
tppubtype = {proceedings}
}
Using a thorough set of environmental data, the Farm to Folk initiative forecasts the most appropriate crops for production by leveraging machine learning methods. This data set includes crucial agricultural indicators, including nitrogen levels, phosphorus content, temperature, humidity, pH levels of water, and other pertinent variables. Using advanced data processing techniques, feature engineering methodologies, and iterative model training procedures, the system attempts to accurately predict the optimal crops to be grown under specific environmental conditions. The efficacy of the predictive model is rigorously evaluated against real-world agricultural scenarios and historical crop yield data. By means of this iterative process, Farm-to-People seeks to provide farmers with a consistent decision-support tool that takes several environmental factors into account, thereby empowering them to make wise crop choice. By facilitating precision agriculture through machine learning-driven predictions, this initiative ultimately aims to preagricultural efficiency, optimize resource utilization, and propromoteustainable farming practices.
16.
Varalatchoumy, M.; Reddy, A. S.; Imambi, S. S.; Ameta, G. K.; Thaiyalnayaki, K.; Logesh, S. K.; Dhivya, S.; Muthulakshmi, K.
Privacy-Preserving Deep Learning Models For Alcoholism Diagnosis Through EEG Data Analysis Using Differential Privacy Mechanisms Journal Article
In: International Journal of Basic and Applied Sciences, vol. 14, pp. 493-503,, 2025, ISBN: 22275053 (ISSN), (0).
@article{463,
title = {Privacy-Preserving Deep Learning Models For Alcoholism Diagnosis Through EEG Data Analysis Using Differential Privacy Mechanisms},
author = {M. Varalatchoumy and A. S. Reddy and S. S. Imambi and G. K. Ameta and K. Thaiyalnayaki and S. K. Logesh and S. Dhivya and K. Muthulakshmi},
url = {https://sciencepubco.com/index.php/IJBAS/article/view/33744},
doi = {10.14419/7gtxb837},
isbn = {22275053 (ISSN)},
year = {2025},
date = {2025-01-01},
journal = {International Journal of Basic and Applied Sciences},
volume = {14},
pages = {493-503,},
publisher = {Science Publishing Corporation Inc.},
abstract = {Alcoholism diagnosis through electroencephalogram (EEG) data analysis offers a promising alternative to traditional methods by identifying specific brain activity patterns associated with alcohol dependency. While deep learning techniques have demonstrated high accuracy in classifying EEG signals, privacy concerns related to sensitive medical data remain prevalent. Ensuring the privacy of patient data is critical for building trust and enabling the adoption of these tools in real-world clinical settings. This study develops deep learning models with enhanced privacy guarantees by incorporating differential privacy mechanisms, including (ϵ, δ)-Differential Privacy and Gaussian Differential Privacy (GDP). We compare their efficacy in preserving data privacy while maintaining model utility. Experiments show that convolutional and long-short-term-memory models optimized with Adam excel in utility and stability. GDP outperforms (ϵ, δ)-DP by requiring less noise, while DP-Adam surpasses DP-SGD in privacy and utility, particularly for fast convergence. Larger datasets further enhance this balance, emphasizing the importance of effective privacy mechanisms and sufficient data. By balancing privacy and utility, this work con-contributes a novel approach to privacy-preserving AI for sensitive health applications, emphasizing scalable models that maintain diagnostic accuracy.},
note = {0},
keywords = {AIML},
pubstate = {published},
tppubtype = {article}
}
Alcoholism diagnosis through electroencephalogram (EEG) data analysis offers a promising alternative to traditional methods by identifying specific brain activity patterns associated with alcohol dependency. While deep learning techniques have demonstrated high accuracy in classifying EEG signals, privacy concerns related to sensitive medical data remain prevalent. Ensuring the privacy of patient data is critical for building trust and enabling the adoption of these tools in real-world clinical settings. This study develops deep learning models with enhanced privacy guarantees by incorporating differential privacy mechanisms, including (ϵ, δ)-Differential Privacy and Gaussian Differential Privacy (GDP). We compare their efficacy in preserving data privacy while maintaining model utility. Experiments show that convolutional and long-short-term-memory models optimized with Adam excel in utility and stability. GDP outperforms (ϵ, δ)-DP by requiring less noise, while DP-Adam surpasses DP-SGD in privacy and utility, particularly for fast convergence. Larger datasets further enhance this balance, emphasizing the importance of effective privacy mechanisms and sufficient data. By balancing privacy and utility, this work con-contributes a novel approach to privacy-preserving AI for sensitive health applications, emphasizing scalable models that maintain diagnostic accuracy.
17.
Revathi, S.; Meesala, S.; Sudha, V.; Priyanka, R.; Manikandakumar, M.; Vigneshwaran, T.
Algorithmic Crypto Trading using EMA Strategy Proceedings
Institute of Electrical and Electronics Engineers Inc., 2025, ISBN: 979-833153519-3 (ISBN), (0).
@proceedings{462,
title = {Algorithmic Crypto Trading using EMA Strategy},
author = {S. Revathi and S. Meesala and V. Sudha and R. Priyanka and M. Manikandakumar and T. Vigneshwaran},
url = {https://ieeexplore.ieee.org/document/11035368},
doi = {10.1109/ICPCSN65854.2025.11035368},
isbn = {979-833153519-3 (ISBN)},
year = {2025},
date = {2025-01-01},
journal = {Proceedings of 5th International Conference on Pervasive Computing and Social Networking, ICPCSN 2025},
pages = {997-1002,},
publisher = {Institute of Electrical and Electronics Engineers Inc.},
abstract = {Algorithmic trading has transformed financial markets by enabling data-driven strategies that enhance efficiency and decision-making. This paper presents a web-based crypto currency trading platform that employs the Exponential Moving Average (EMA) strategy for automated trade execution, market trend analysis, and portfolio tracking. The platform integrates key performance metrics, including win rate, average profit per trade, risk-reward ratio, and profit factor to assess trading effectiveness. Notably, EMA-based trading achieves the highest profit factor of 3.5 which outperformed deep learning and manual trading by 9.37% and 133%, respectively. Additionally, EMA exhibits a strong win rate of 60%, compared to 65% for deep learning and 40% for manual trading, while maintaining a balanced risk-reward ratio of 2.2. The system features live data visualization, customizable watchlists, and automated trading workflows, providing traders with actionable insights with minimized human error. Performance evaluation indicates that EMA offers a superior trade-off between profitability and risk management, making it a robust and adaptable solution for navigating cryptocurrency markets. This work bridges the gap between manual trading and advanced algorithmic strategies, delivering a user-friendly and efficient trading framework.},
note = {0},
keywords = {AIML},
pubstate = {published},
tppubtype = {proceedings}
}
Algorithmic trading has transformed financial markets by enabling data-driven strategies that enhance efficiency and decision-making. This paper presents a web-based crypto currency trading platform that employs the Exponential Moving Average (EMA) strategy for automated trade execution, market trend analysis, and portfolio tracking. The platform integrates key performance metrics, including win rate, average profit per trade, risk-reward ratio, and profit factor to assess trading effectiveness. Notably, EMA-based trading achieves the highest profit factor of 3.5 which outperformed deep learning and manual trading by 9.37% and 133%, respectively. Additionally, EMA exhibits a strong win rate of 60%, compared to 65% for deep learning and 40% for manual trading, while maintaining a balanced risk-reward ratio of 2.2. The system features live data visualization, customizable watchlists, and automated trading workflows, providing traders with actionable insights with minimized human error. Performance evaluation indicates that EMA offers a superior trade-off between profitability and risk management, making it a robust and adaptable solution for navigating cryptocurrency markets. This work bridges the gap between manual trading and advanced algorithmic strategies, delivering a user-friendly and efficient trading framework.
18.
Ramya, R.; Premalatha, R.; Prasad, A. R.; Pudi, A.; Raju, C. V. V. N.; Varalatchoumy, M.
Smart solutions for electric vehicles using AI in mobility and infrastructure Book Chapter
In: pp. 207-226,, IGI Global, 2025, ISBN: 979-836937665-2 (ISBN); 979-836937663-8 (ISBN), (0).
@inbook{422,
title = {Smart solutions for electric vehicles using AI in mobility and infrastructure},
author = {R. Ramya and R. Premalatha and A. R. Prasad and A. Pudi and C. V. V. N. Raju and M. Varalatchoumy},
url = {https://www.igi-global.com/gateway/chapter/376208},
doi = {10.4018/979-8-3693-7663-8.ch010},
isbn = {979-836937665-2 (ISBN); 979-836937663-8 (ISBN)},
year = {2025},
date = {2025-01-01},
journal = {Innovations in Power Systems and Applications},
pages = {207-226,},
publisher = {IGI Global},
abstract = {This chapter explores the transformative role of artificial intelligence in enhancing the mobility and infrastructure of electric vehicles. As electric vehicle adoption surges, intelligent systems must be optimized for vehicle performance, charging networks, and urban mobility solutions. The integration of machine learning, predictive analytics, and deep learning technologies has significantly transformed the way electric vehicles interact with charging stations, road networks, and users. AI enhances energy efficiency, predicts maintenance needs, and enables autonomous driving for EVs. AI- based infrastructure management systems optimize location planning, real- time monitoring, and load balancing for charging stations. The text discusses the significant applications of AI in route optimization, energy management, and smart city integration. The chapter discusses how AI enhances electric mobility efficiency and accelerates the transition towards sustainable transportation ecosystems. © 2025, IGI Global Scientific Publishing.},
note = {0},
keywords = {AIML},
pubstate = {published},
tppubtype = {inbook}
}
This chapter explores the transformative role of artificial intelligence in enhancing the mobility and infrastructure of electric vehicles. As electric vehicle adoption surges, intelligent systems must be optimized for vehicle performance, charging networks, and urban mobility solutions. The integration of machine learning, predictive analytics, and deep learning technologies has significantly transformed the way electric vehicles interact with charging stations, road networks, and users. AI enhances energy efficiency, predicts maintenance needs, and enables autonomous driving for EVs. AI- based infrastructure management systems optimize location planning, real- time monitoring, and load balancing for charging stations. The text discusses the significant applications of AI in route optimization, energy management, and smart city integration. The chapter discusses how AI enhances electric mobility efficiency and accelerates the transition towards sustainable transportation ecosystems. © 2025, IGI Global Scientific Publishing.
19.
Komala, C. R.; Varalatchoumy, M.; Kulkarni, M.; Setty, S.; Kousar, H.; Boopathi, S.
IoT-driven automation systems for hydroponic agriculture Book Chapter
In: pp. 357-378,, IGI Global, 2025, ISBN: 979-836937114-5 (ISBN); 979-836937112-1 (ISBN), (0).
@inbook{411,
title = {IoT-driven automation systems for hydroponic agriculture},
author = {C. R. Komala and M. Varalatchoumy and M. Kulkarni and S. Setty and H. Kousar and S. Boopathi},
url = {https://www.igi-global.com/gateway/chapter/374518},
doi = {10.4018/979-8-3693-7112-1.ch017},
isbn = {979-836937114-5 (ISBN); 979-836937112-1 (ISBN)},
year = {2025},
date = {2025-01-01},
journal = {Integrating Artificial Intelligence Into the Energy Sector},
pages = {357-378,},
publisher = {IGI Global},
abstract = {The paper develops an analysis in terms of the transformational impacts of IoT-driven automation systems in hydroponic agriculture. Sensors, actuators, and data analytics through IoT enable the creation of smart responsive environments for hydroponic systems. IoT automation enables real-time monitoring and control of variables relating to nutrient levels, pH, temperature, and humidity for optimally maximized growth and resource utilization. The architectural framework of IoT systems in hydroponics is explained with regard to sensor deployment for real-time data, as well as utilization of machine learning algorithms for predictive analytics, in addition to case studies illustrating improved yields, reduction in water and nutrient utilization, and lower operational cost. And finally, the chapter will conclude with insights into future trends, which include further advancements in AI and robotics, leading to further innovations in automated hydroponic farming systems.},
note = {0},
keywords = {AIML},
pubstate = {published},
tppubtype = {inbook}
}
The paper develops an analysis in terms of the transformational impacts of IoT-driven automation systems in hydroponic agriculture. Sensors, actuators, and data analytics through IoT enable the creation of smart responsive environments for hydroponic systems. IoT automation enables real-time monitoring and control of variables relating to nutrient levels, pH, temperature, and humidity for optimally maximized growth and resource utilization. The architectural framework of IoT systems in hydroponics is explained with regard to sensor deployment for real-time data, as well as utilization of machine learning algorithms for predictive analytics, in addition to case studies illustrating improved yields, reduction in water and nutrient utilization, and lower operational cost. And finally, the chapter will conclude with insights into future trends, which include further advancements in AI and robotics, leading to further innovations in automated hydroponic farming systems.
20.
Kavitha, D.; Ramkumar, Siva; Arun, M.; Baig, R. U.; Sivaramkrishnan, M.
Design Thinking in Electrical Engineering: A Pathway to Human-Centered Innovations Proceedings
Institute of Electrical and Electronics Engineers Inc., 2025, ISBN: 979-833150967-5 (ISBN), (0).
@proceedings{410,
title = {Design Thinking in Electrical Engineering: A Pathway to Human-Centered Innovations},
author = {D. Kavitha and Siva Ramkumar and M. Arun and R. U. Baig and M. Sivaramkrishnan},
url = {https://ieeexplore.ieee.org/document/10940337},
doi = {10.1109/ICEARS64219.2025.10940337},
isbn = {979-833150967-5 (ISBN)},
year = {2025},
date = {2025-01-01},
journal = {3rd International Conference on Electronics and Renewable Systems, ICEARS 2025 - Proceedings},
pages = {288-295,},
publisher = {Institute of Electrical and Electronics Engineers Inc.},
abstract = {Design thinking is a useful technique in many technical domains since it is a novel methodology that emphasizes comprehending and meeting customer demands. Using design thinking in electrical engineering provides a fresh approach to creating user-focused solutions in a field where technological complexity is constantly changing. In order to foster creativity and useful problem-solving, this study explores how applying design thinking concepts might transform electrical engineering's conventional procedures. Engineers may produce more efficient and user-friendly designs by utilizing crucial phases like empathy, problem identification, ideation, prototyping, and iterative testing. In order to demonstrate how this strategy results in more flexible and sustainable solutions, the discussion cites particular instances, such as developments in consumer electronics, electric power systems, and renewable energy. The study also emphasizes how teamwork and ongoing iteration foster creativity while maintaining designs' alignment with user requirements and societal impact.},
note = {0},
keywords = {AIML},
pubstate = {published},
tppubtype = {proceedings}
}
Design thinking is a useful technique in many technical domains since it is a novel methodology that emphasizes comprehending and meeting customer demands. Using design thinking in electrical engineering provides a fresh approach to creating user-focused solutions in a field where technological complexity is constantly changing. In order to foster creativity and useful problem-solving, this study explores how applying design thinking concepts might transform electrical engineering's conventional procedures. Engineers may produce more efficient and user-friendly designs by utilizing crucial phases like empathy, problem identification, ideation, prototyping, and iterative testing. In order to demonstrate how this strategy results in more flexible and sustainable solutions, the discussion cites particular instances, such as developments in consumer electronics, electric power systems, and renewable energy. The study also emphasizes how teamwork and ongoing iteration foster creativity while maintaining designs' alignment with user requirements and societal impact.
21.
Sivaramkrishnan, M.; Ramkumar, Siva; Kannaiyan, M.; Kanan, M.; Baig, R. U.; Giri, J.
Institute of Electrical and Electronics Engineers Inc., 2025, ISBN: 979-833152392-3 (ISBN), (0).
@proceedings{409,
title = {EEG-Based Emotion Recognition Using Morlet Dual-Level Wavelet Contextual Neural Network with Snow Geese Algorithm},
author = {M. Sivaramkrishnan and Siva Ramkumar and M. Kannaiyan and M. Kanan and R. U. Baig and J. Giri},
url = {https://ieeexplore.ieee.org/document/10933301},
doi = {10.1109/ICSADL65848.2025.10933301},
isbn = {979-833152392-3 (ISBN)},
year = {2025},
date = {2025-01-01},
journal = {4th International Conference on Sentiment Analysis and Deep Learning, ICSADL 2025 - Proceedings},
pages = {1578-1584,},
publisher = {Institute of Electrical and Electronics Engineers Inc.},
abstract = {EEG-based Emotion Recognition (ER) works by analyzing brainwave patterns from EEG signals to understand and identify a person's emotional state. The problem of emotion identification from EEG signals remains challenging because brain activity presents both non-linear dynamic properties and individual-specific EEG patterns along with biological noise. Traditional approaches to emotion detection become inefficient due to EEG signal complexity thereby complicating accuracy and reliability measures. To tackle the challenges in EEG-based emotion recognition, this study presents a new approach called the Morlet Dual-level Wavelet Contextual Neural Network with Snow Geese Algorithm (MorDWCNNet+SGA). The method uses the SEED dataset, starting with preprocessing through the Observability-Constrained Resampling-Free Cubature Kalman Filter (OCRCKF) to enhance important EEG patterns. For feature extraction, the Multi-Discrete Wavelet Transform (MDWT) is applied to capture key EEG characteristics. Emotion classification is then performed using the Morlet Dual-level Wavelet Contextual Neural Network (MorDWCNNet), which is further optimized by the Snow Geese Algorithm (SGA) to improve accuracy. Implemented in Python and tested on the SEED dataset, the MorDWCNNet+SGA framework achieves remarkable results, with 99.9% accuracy and 99.3% sensitivity, significantly outperforming existing methods. The outcomes of the proposed method achieved high accuracy in effectively distinguishing emotional states through brainwave patterns. This method demonstrates how advanced techniques can combine synergistically to enhance ER performance resulting in improved real-world application accuracy and speed.},
note = {0},
keywords = {AIML},
pubstate = {published},
tppubtype = {proceedings}
}
EEG-based Emotion Recognition (ER) works by analyzing brainwave patterns from EEG signals to understand and identify a person's emotional state. The problem of emotion identification from EEG signals remains challenging because brain activity presents both non-linear dynamic properties and individual-specific EEG patterns along with biological noise. Traditional approaches to emotion detection become inefficient due to EEG signal complexity thereby complicating accuracy and reliability measures. To tackle the challenges in EEG-based emotion recognition, this study presents a new approach called the Morlet Dual-level Wavelet Contextual Neural Network with Snow Geese Algorithm (MorDWCNNet+SGA). The method uses the SEED dataset, starting with preprocessing through the Observability-Constrained Resampling-Free Cubature Kalman Filter (OCRCKF) to enhance important EEG patterns. For feature extraction, the Multi-Discrete Wavelet Transform (MDWT) is applied to capture key EEG characteristics. Emotion classification is then performed using the Morlet Dual-level Wavelet Contextual Neural Network (MorDWCNNet), which is further optimized by the Snow Geese Algorithm (SGA) to improve accuracy. Implemented in Python and tested on the SEED dataset, the MorDWCNNet+SGA framework achieves remarkable results, with 99.9% accuracy and 99.3% sensitivity, significantly outperforming existing methods. The outcomes of the proposed method achieved high accuracy in effectively distinguishing emotional states through brainwave patterns. This method demonstrates how advanced techniques can combine synergistically to enhance ER performance resulting in improved real-world application accuracy and speed.
22.
Ramkumar, M. S.; Nivetha, R.; Gopan, G.; Kanan, M.; Baig, R. U.; Giri, J.
Institute of Electrical and Electronics Engineers Inc., 2025, ISBN: 979-833152392-3 (ISBN), (0).
@proceedings{408,
title = {Sustainable Crop Yield Forecasting with the Dual-Branch Geometric Progressive Feedback Cosine Convolutional Network Model},
author = {M. S. Ramkumar and R. Nivetha and G. Gopan and M. Kanan and R. U. Baig and J. Giri},
url = {https://ieeexplore.ieee.org/document/10933059},
doi = {10.1109/ICSADL65848.2025.10933059},
isbn = {979-833152392-3 (ISBN)},
year = {2025},
date = {2025-01-01},
journal = {4th International Conference on Sentiment Analysis and Deep Learning, ICSADL 2025 - Proceedings},
pages = {1592-1598,},
publisher = {Institute of Electrical and Electronics Engineers Inc.},
abstract = {Using deep learning algorithms to extract important agricultural traits has made crop production prediction based on environmental, soil, water, and crop parameters an important area of research. But conventional approaches have trouble creating a straight linear or non-linear mapping between yield values and raw data, and the quality of the characteristics that are extracted has a significant impact on how well they work. Deep reinforcement learning builds a strong prediction framework by combining the intelligence of deep learning with reinforcement learning to get beyond these drawbacks. In order to anticipate agricultural yields accurately and sustainably, this work aims to create a Dual-Branch Geometric Progressive Feedback Cosine Convolutional Network with Skill Optimization Algorithm (DB-GPFCNet-SOA). Robust Double-Weighted Guided Filtering (RDWGF) is used for data preprocessing in order to minimize noise and improve important components. To ensure a complete prediction model, the N-Branch retrieves local neighbor-based information while the C-Branch records global geometric properties. Moreover, the accuracy is enhanced by the SOA, which enhances loss parameters. The suggested model has RMSE of 0.25 and MAE of 0.15 with exceptional accuracy of 99.2%, precision of 99.1%, recall of 99%, and F1-score of 99%. These results validate the performance of the model, which renders it a suitable approach for real-time agricultural applications.},
note = {0},
keywords = {AIML},
pubstate = {published},
tppubtype = {proceedings}
}
Using deep learning algorithms to extract important agricultural traits has made crop production prediction based on environmental, soil, water, and crop parameters an important area of research. But conventional approaches have trouble creating a straight linear or non-linear mapping between yield values and raw data, and the quality of the characteristics that are extracted has a significant impact on how well they work. Deep reinforcement learning builds a strong prediction framework by combining the intelligence of deep learning with reinforcement learning to get beyond these drawbacks. In order to anticipate agricultural yields accurately and sustainably, this work aims to create a Dual-Branch Geometric Progressive Feedback Cosine Convolutional Network with Skill Optimization Algorithm (DB-GPFCNet-SOA). Robust Double-Weighted Guided Filtering (RDWGF) is used for data preprocessing in order to minimize noise and improve important components. To ensure a complete prediction model, the N-Branch retrieves local neighbor-based information while the C-Branch records global geometric properties. Moreover, the accuracy is enhanced by the SOA, which enhances loss parameters. The suggested model has RMSE of 0.25 and MAE of 0.15 with exceptional accuracy of 99.2%, precision of 99.1%, recall of 99%, and F1-score of 99%. These results validate the performance of the model, which renders it a suitable approach for real-time agricultural applications.
23.
Ramkumar, M. S.; Saranya, N.; Gopan, G.; Baig, R. U.; Daniel, Josha; Babu, M.
Institute of Electrical and Electronics Engineers Inc., 2025, ISBN: 979-833150967-5 (ISBN), (0).
@proceedings{407,
title = {Giant Trevally Optimizer (GTO): Enhancing HVDC Transmission Capacity with Optimized Fault Current Limiters and HVDC Circuit Breaker Parameters},
author = {M. S. Ramkumar and N. Saranya and G. Gopan and R. U. Baig and Josha Daniel and M. Babu},
url = {https://ieeexplore.ieee.org/document/10940142},
doi = {10.1109/ICEARS64219.2025.10940142},
isbn = {979-833150967-5 (ISBN)},
year = {2025},
date = {2025-01-01},
journal = {3rd International Conference on Electronics and Renewable Systems, ICEARS 2025 - Proceedings},
pages = {268-274,},
publisher = {Institute of Electrical and Electronics Engineers Inc.},
abstract = {The multi-terminal HVDC system relies heavily on circuit breakers (CBs) and fault current limiters (FCLs) for protection and performance reliability. This research describes a new Giant Trevally optimizer-based strategy for improving the performance of 4-terminal HVDC systems by optimizing crucial circuit breakers and fault current limiters. The proposed method uses time domain estimation to correctly capture fault dynamics, such as traveling waves and reflection coefficients and finds the worst-case fault locations. The multi-objective optimization seeks to optimize the overcurrent, overvoltage handling, fault clearance time, and energy absorption in arresters. This technology provides an effective, dependable protection solution that reduces fault recovery time and increases system stability. The output pulse of the HVDC system shows very low THD results 1.98% for the minimum load condition, 0.87% for median load, and 0.04% for peak load conditions. The results established the simulation of the model in a MATLAB environment.},
note = {0},
keywords = {AIML},
pubstate = {published},
tppubtype = {proceedings}
}
The multi-terminal HVDC system relies heavily on circuit breakers (CBs) and fault current limiters (FCLs) for protection and performance reliability. This research describes a new Giant Trevally optimizer-based strategy for improving the performance of 4-terminal HVDC systems by optimizing crucial circuit breakers and fault current limiters. The proposed method uses time domain estimation to correctly capture fault dynamics, such as traveling waves and reflection coefficients and finds the worst-case fault locations. The multi-objective optimization seeks to optimize the overcurrent, overvoltage handling, fault clearance time, and energy absorption in arresters. This technology provides an effective, dependable protection solution that reduces fault recovery time and increases system stability. The output pulse of the HVDC system shows very low THD results 1.98% for the minimum load condition, 0.87% for median load, and 0.04% for peak load conditions. The results established the simulation of the model in a MATLAB environment.
24.
Maheswari, P.; Raja, P.; Karkee, M.; Raja, M.; Baig, R. U.; Trung, K. T.; Hoang, V. T.
Performance analysis of modified DeepLabv3+ architecture for fruit detection and localization in apple orchards Journal Article
In: Smart Agricultural Technology, vol. 10, 2025, ISBN: 27723755 (ISSN), (0).
@article{352,
title = {Performance analysis of modified DeepLabv3+ architecture for fruit detection and localization in apple orchards},
author = {P. Maheswari and P. Raja and M. Karkee and M. Raja and R. U. Baig and K. T. Trung and V. T. Hoang},
url = {https://www.sciencedirect.com/science/article/pii/S2772375524003332?via%3Dihub},
doi = {10.1016/j.atech.2024.100729},
isbn = {27723755 (ISSN)},
year = {2025},
date = {2025-01-01},
journal = {Smart Agricultural Technology},
volume = {10},
publisher = {Elsevier B.V.},
abstract = {Deep learning plays an important role in automating various operations in fruit crop production including irrigation, nutrition management, yield estimation and harvesting. Yield estimation is essential in fruit crop production as it helps farmers optimize cultivation, harvesting, logistics and marketing operations. Furthermore, fruit detection and localization is a very important step in the development of an automated fruit harvesting system. Hence, an intelligent system was proposed in this study for apple fruit detection and localization using modified DeepLabv3+, semantic segmentation based architecture. The finetuned customizations (such as modifying the activation function, optimization technique and loss function) were performed in the original architecture of DeepLabv3+ and its performance was analyzed. The modified model was trained with the training dataset of 2600 apple tree images. Images were split into 80 % of training and 20 % of validation. The modified architecture was also compared with the other variants of DeepLabv3+ architectures. After training, the model was tested with the unobserved test dataset of 101 images. The test results demonstrated the Mean Accuracy (MAcc) of 98.58 % and the Mean Intersection over Union (MIoU) of 96.66 % without compromising the inference time (i.e., 15 ms). The proposed model revealed the improved results than the original model which attained a MAcc of 92.12 % and MIoU of 88.94 % for the same dataset with the inference time of 40 ms. To ascertain further, the modified model was compared with other single stage detectors, including Fully Convolutional Network (FCN) and U-Net. FCN attained a MAccandMIoU of 77.5 % and 77.27 %, respectively whereas U-Net resulted a MAcc and MIoU of 83.95 % and 81.09 %, respectively. Results demonstrated that the modified DeepLabv3+ with ResNet18 is capable of detecting the apple fruits by mitigating the effects of class imbalance which is the major drawback in single stage detectors. Further, better detection and localization of apple fruits can lead to the precise picking by the robotic system.},
note = {0},
keywords = {AIML},
pubstate = {published},
tppubtype = {article}
}
Deep learning plays an important role in automating various operations in fruit crop production including irrigation, nutrition management, yield estimation and harvesting. Yield estimation is essential in fruit crop production as it helps farmers optimize cultivation, harvesting, logistics and marketing operations. Furthermore, fruit detection and localization is a very important step in the development of an automated fruit harvesting system. Hence, an intelligent system was proposed in this study for apple fruit detection and localization using modified DeepLabv3+, semantic segmentation based architecture. The finetuned customizations (such as modifying the activation function, optimization technique and loss function) were performed in the original architecture of DeepLabv3+ and its performance was analyzed. The modified model was trained with the training dataset of 2600 apple tree images. Images were split into 80 % of training and 20 % of validation. The modified architecture was also compared with the other variants of DeepLabv3+ architectures. After training, the model was tested with the unobserved test dataset of 101 images. The test results demonstrated the Mean Accuracy (MAcc) of 98.58 % and the Mean Intersection over Union (MIoU) of 96.66 % without compromising the inference time (i.e., 15 ms). The proposed model revealed the improved results than the original model which attained a MAcc of 92.12 % and MIoU of 88.94 % for the same dataset with the inference time of 40 ms. To ascertain further, the modified model was compared with other single stage detectors, including Fully Convolutional Network (FCN) and U-Net. FCN attained a MAccandMIoU of 77.5 % and 77.27 %, respectively whereas U-Net resulted a MAcc and MIoU of 83.95 % and 81.09 %, respectively. Results demonstrated that the modified DeepLabv3+ with ResNet18 is capable of detecting the apple fruits by mitigating the effects of class imbalance which is the major drawback in single stage detectors. Further, better detection and localization of apple fruits can lead to the precise picking by the robotic system.
25.
Hayath, S.; Geetha, R.; Latha, C. R.; Mahalakshmi, V.; Anusha, K. V.
Smart Watches and Pulse Bands with Built-in Oximeter Proceedings
Institute of Electrical and Electronics Engineers Inc., 2024, ISBN: 979-835034387-8 (ISBN), (0).
@proceedings{464,
title = {Smart Watches and Pulse Bands with Built-in Oximeter},
author = {S. Hayath and R. Geetha and C. R. Latha and V. Mahalakshmi and K. V. Anusha},
url = {https://ieeexplore.ieee.org/document/11007413},
doi = {10.1109/ISML60050.2024.11007413},
isbn = {979-835034387-8 (ISBN)},
year = {2024},
date = {2024-01-01},
journal = {ISML 2024 - Intelligent Systems and Machine Learning Conference},
pages = {101-105,},
publisher = {Institute of Electrical and Electronics Engineers Inc.},
abstract = {In the present day we live in, a catastrophic virus COVID-19 has drastically changed the day to day living of a human being. The virus causes hypoxemia (shortness of breath), sometimes even severe cases. Hypoxemia is caused due to low oxygen saturation (SpO2) levels and in order to monitor SpO2 levels we utilize the method of Pulse Oximetry. It has played a major role in monitoring oxygen levels in a continuous and accurate manner. Patients who are diagnosed during the early stages with the virus are able to monitor their SpO2 levels at home and take the prescribed medicines. This reduced the need for hospital beds and chaos during the pandemic as patients were able to self-isolate and monitor themselves. The principle that Oximetry is contingent on is light reflectance. The fingertip model is the most common technique used to display the blood oxygenation level. However, the finger model is not as comfortable and does not promote continuous monitoring of SpO2 levels. Factors such as nail polish can play a role in the accuracy of the reading. The proposed design is a non-invasive medical device that can measure the oxygen saturation level in a person's blood. Wrist wearing is preferable to finger models as it is a comfortable site for measuring than that of the latter, therefore enabling continuous monitoring of oxygen saturation level. Most devices in application today are inaccurate and used for non-clinical use. Therefore, it is aimed to achieve better accuracy by avoiding readings that are influenced by factors such as motion artifacts by utilizing sensors that are appropriate through an accelerometer and gyroscope. The sensors have been utilized to its maximum potential as the model also detects heart rate levels. Providing device connectivity to the proposed wrist pulse oximeter through a Wi- Fi module has advanced its functionality by transmitting notifications that alert an individual with the levels of oxygen saturation, fall detection and heart rate levels.},
note = {0},
keywords = {AIML},
pubstate = {published},
tppubtype = {proceedings}
}
In the present day we live in, a catastrophic virus COVID-19 has drastically changed the day to day living of a human being. The virus causes hypoxemia (shortness of breath), sometimes even severe cases. Hypoxemia is caused due to low oxygen saturation (SpO2) levels and in order to monitor SpO2 levels we utilize the method of Pulse Oximetry. It has played a major role in monitoring oxygen levels in a continuous and accurate manner. Patients who are diagnosed during the early stages with the virus are able to monitor their SpO2 levels at home and take the prescribed medicines. This reduced the need for hospital beds and chaos during the pandemic as patients were able to self-isolate and monitor themselves. The principle that Oximetry is contingent on is light reflectance. The fingertip model is the most common technique used to display the blood oxygenation level. However, the finger model is not as comfortable and does not promote continuous monitoring of SpO2 levels. Factors such as nail polish can play a role in the accuracy of the reading. The proposed design is a non-invasive medical device that can measure the oxygen saturation level in a person's blood. Wrist wearing is preferable to finger models as it is a comfortable site for measuring than that of the latter, therefore enabling continuous monitoring of oxygen saturation level. Most devices in application today are inaccurate and used for non-clinical use. Therefore, it is aimed to achieve better accuracy by avoiding readings that are influenced by factors such as motion artifacts by utilizing sensors that are appropriate through an accelerometer and gyroscope. The sensors have been utilized to its maximum potential as the model also detects heart rate levels. Providing device connectivity to the proposed wrist pulse oximeter through a Wi- Fi module has advanced its functionality by transmitting notifications that alert an individual with the levels of oxygen saturation, fall detection and heart rate levels.
26.
Naveen, S.; Sanketh, G.; Jayanthi, M.; Shreya, M.; Kannadaguli, P.
RouteRover: AI-Enabled Traffic Congestion Prediction and Route Optimization for Indian Urbanites Proceedings
2024.
@proceedings{402,
title = {RouteRover: AI-Enabled Traffic Congestion Prediction and Route Optimization for Indian Urbanites},
author = {S. Naveen and G. Sanketh and M. Jayanthi and M. Shreya and P. Kannadaguli},
url = {https://ieeexplore.ieee.org/document/10895019},
doi = {10.1109/ICRASET63057.2024.10895019},
year = {2024},
date = {2024-01-01},
journal = {2024 International Conference on Recent Advances in Science and Engineering Technology (ICRASET)},
pages = {1-5,},
abstract = {Here in this paper, a model that has been proposed for predicting and visually representing congestion of traffic using machine learning as well as deep learning. This initiative with the purpose of examining traffic data to obtain insights for traffic regulation and planning. The process initiates with an initial exploration of data through an Exploratory Data Analysis (EDA) and to perceive the distribution and patterns within the dataset. Descriptive statistics, correlation analysis, and visual aids are employed for this purpose. Following the EDA phase, tactics for manifold learning are utilized to simplify the dataset while retaining crucial information. Identifying significant features influencing the prediction of traffic volume is facilitated by methods for feature importance and selection. The obtained features are rigorously evaluated to comprehend traffic patterns and traffic jam levels. Visual representations capture the distribution of traffic volume based on factors like time, day, and month, assisting in pinpointing congestion hotspots and optimizing traffic flow. This project underscores the powerful impact of data analysis and feature extraction in offering practical insights for traffic regulation strategies.},
keywords = {AIML},
pubstate = {published},
tppubtype = {proceedings}
}
Here in this paper, a model that has been proposed for predicting and visually representing congestion of traffic using machine learning as well as deep learning. This initiative with the purpose of examining traffic data to obtain insights for traffic regulation and planning. The process initiates with an initial exploration of data through an Exploratory Data Analysis (EDA) and to perceive the distribution and patterns within the dataset. Descriptive statistics, correlation analysis, and visual aids are employed for this purpose. Following the EDA phase, tactics for manifold learning are utilized to simplify the dataset while retaining crucial information. Identifying significant features influencing the prediction of traffic volume is facilitated by methods for feature importance and selection. The obtained features are rigorously evaluated to comprehend traffic patterns and traffic jam levels. Visual representations capture the distribution of traffic volume based on factors like time, day, and month, assisting in pinpointing congestion hotspots and optimizing traffic flow. This project underscores the powerful impact of data analysis and feature extraction in offering practical insights for traffic regulation strategies.
27.
Prema, S.; Varalatchoumy, M.; Nirmaladevi, G.; Vijayakumar, S.; Kayalvili, S.; Rajendiran, M.; Premanand, R.; Vijayan, V.
American Institute of Physics, vol. 3193, 2024, ISBN: 0094243X (ISSN), (0).
@proceedings{347,
title = {Advancements in machine learning for recommender systems: A focus on NNMFC and particle swarm optimization techniques},
author = {S. Prema and M. Varalatchoumy and G. Nirmaladevi and S. Vijayakumar and S. Kayalvili and M. Rajendiran and R. Premanand and V. Vijayan},
url = {https://pubs.aip.org/aip/acp/article-abstract/3193/1/020019/3319617/Advancements-in-machine-learning-for-recommender?redirectedFrom=fulltext},
doi = {10.1063/5.0235519},
isbn = {0094243X (ISSN)},
year = {2024},
date = {2024-01-01},
journal = {AIP Conference Proceedings},
volume = {3193},
publisher = {American Institute of Physics},
abstract = {Through the use of interest models, the Recommender System assists users in discovering content that is relevant to them. In order to make product suggestions based on past purchases, content-based recommender systems do not require user ratings. These systems are the subject of this study. Now these systems can profile products and customers using machine learning. Together with Non-Negative Matrix Factorization Clustering (NNMFC), we present a new approach to collaborative learning for online video sites. The research utilizes a sliding window clustering approach that relies on Particle Swarm Optimization (PSO) and gradient descent. We utilized three approaches to assess the model's performance: sliding window PSO (SWPSO), sliding window gradient descent and gradient descent. The Root Mean Square Error (RMSE) was calculated for each. Outperforming current state-of-the-art methods like UPCSim, K-Mean, and Collaborative Filtering, the suggested work's result analysis attained the lowest RMSE of 1.02. With a significant improvement of 10.2% over previous techniques, the Sliding Window PSO was shown to be the most effective.},
note = {0},
keywords = {AIML},
pubstate = {published},
tppubtype = {proceedings}
}
Through the use of interest models, the Recommender System assists users in discovering content that is relevant to them. In order to make product suggestions based on past purchases, content-based recommender systems do not require user ratings. These systems are the subject of this study. Now these systems can profile products and customers using machine learning. Together with Non-Negative Matrix Factorization Clustering (NNMFC), we present a new approach to collaborative learning for online video sites. The research utilizes a sliding window clustering approach that relies on Particle Swarm Optimization (PSO) and gradient descent. We utilized three approaches to assess the model's performance: sliding window PSO (SWPSO), sliding window gradient descent and gradient descent. The Root Mean Square Error (RMSE) was calculated for each. Outperforming current state-of-the-art methods like UPCSim, K-Mean, and Collaborative Filtering, the suggested work's result analysis attained the lowest RMSE of 1.02. With a significant improvement of 10.2% over previous techniques, the Sliding Window PSO was shown to be the most effective.
28.
Nagaraj, T.; Channarayappa, R. K.
An efficient security framework for intrusion detection and prevention in internet-of-things using machine learning technique Journal Article
In: International Journal of Electrical and Computer Engineering, vol. 14, pp. 2313-2321,, 2024, ISBN: 20888708 (ISSN), (2).
@article{82,
title = {An efficient security framework for intrusion detection and prevention in internet-of-things using machine learning technique},
author = {T. Nagaraj and R. K. Channarayappa},
doi = {10.11591/ijece.v14i2.pp2313-2321},
isbn = {20888708 (ISSN)},
year = {2024},
date = {2024-01-01},
journal = {International Journal of Electrical and Computer Engineering},
volume = {14},
pages = {2313-2321,},
publisher = {Institute of Advanced Engineering and Science},
abstract = {Over the past few years, the internet of things (IoT) has advanced to connect billions of smart devices to improve quality of life. However, anomalies or malicious intrusions pose several security loopholes, leading to performance degradation and threat to data security in IoT operations. Thereby, IoT security systems must keep an eye on and restrict unwanted events from occurring in the IoT network. Recently, various technical solutions based on machine learning (ML) models have been derived towards identifying and restricting unwanted events in IoT. However, most ML-based approaches are prone to miss-classification due to inappropriate feature selection. Additionally, most ML approaches applied to intrusion detection and prevention consider supervised learning, which requires a large amount of labeled data to be trained. Consequently, such complex datasets are impossible to source in a large network like IoT. To address this problem, this proposed study introduces an efficient learning mechanism to strengthen the IoT security aspects. The proposed algorithm incorporates supervised and unsupervised approaches to improve the learning models for intrusion detection and mitigation. Compared with the related works, the experimental outcome shows that the model performs well in a benchmark dataset. It accomplishes an improved detection accuracy of approximately 99.21%. © 2024 Institute of Advanced Engineering and Science. All rights reserved.},
note = {2},
keywords = {AIML},
pubstate = {published},
tppubtype = {article}
}
Over the past few years, the internet of things (IoT) has advanced to connect billions of smart devices to improve quality of life. However, anomalies or malicious intrusions pose several security loopholes, leading to performance degradation and threat to data security in IoT operations. Thereby, IoT security systems must keep an eye on and restrict unwanted events from occurring in the IoT network. Recently, various technical solutions based on machine learning (ML) models have been derived towards identifying and restricting unwanted events in IoT. However, most ML-based approaches are prone to miss-classification due to inappropriate feature selection. Additionally, most ML approaches applied to intrusion detection and prevention consider supervised learning, which requires a large amount of labeled data to be trained. Consequently, such complex datasets are impossible to source in a large network like IoT. To address this problem, this proposed study introduces an efficient learning mechanism to strengthen the IoT security aspects. The proposed algorithm incorporates supervised and unsupervised approaches to improve the learning models for intrusion detection and mitigation. Compared with the related works, the experimental outcome shows that the model performs well in a benchmark dataset. It accomplishes an improved detection accuracy of approximately 99.21%. © 2024 Institute of Advanced Engineering and Science. All rights reserved.
29.
Aaditya, P.; Manu, R.; Asif, M.; Jayanthi, M. G.; Kannadaguli, P.
Comparative Analysis of Random Forest and CNN for Surgery Mortality Prediction Proceedings
Institute of Electrical and Electronics Engineers Inc., 2023, ISBN: 979-835030082-6 (ISBN), (0).
@proceedings{58,
title = {Comparative Analysis of Random Forest and CNN for Surgery Mortality Prediction},
author = {P. Aaditya and R. Manu and M. Asif and M. G. Jayanthi and P. Kannadaguli},
doi = {10.1109/NMITCON58196.2023.10276379},
isbn = {979-835030082-6 (ISBN)},
year = {2023},
date = {2023-01-01},
journal = {2023 International Conference on Network, Multimedia and Information Technology, NMITCON 2023},
publisher = {Institute of Electrical and Electronics Engineers Inc.},
abstract = {Accurate surgical outcome forecasting is essential for patient safety, healthcare resource allocation, and healthcare professional's capacity to make well-informed choices. To address this, a paper has been initiated to create a machine learning (ML) and deep learning (DL) model that can effectively determine the likelihood of a patient surviving a surgery. The main goal of this paper is to compare the performance of both models and determine which one yields superior overall results and develop something that can help the healthcare industry in making judgements. To conduct this research, a comprehensive dataset comprising approximately 50,000 surgical cases from various hospitals has been acquired. The paper successfully developed and evaluated various models. The best-performing model was 'Random Forest', which achieved an accuracy of 92.51%. The deep learning models also showed useful results, with the best model achieving an accuracy of 91.3%. The process of fine-tuning the DL model has yielded a significant improvement in accuracy, increasing it from 91.3% to 93.4%. © 2023 IEEE.},
note = {0},
keywords = {AIML},
pubstate = {published},
tppubtype = {proceedings}
}
Accurate surgical outcome forecasting is essential for patient safety, healthcare resource allocation, and healthcare professional's capacity to make well-informed choices. To address this, a paper has been initiated to create a machine learning (ML) and deep learning (DL) model that can effectively determine the likelihood of a patient surviving a surgery. The main goal of this paper is to compare the performance of both models and determine which one yields superior overall results and develop something that can help the healthcare industry in making judgements. To conduct this research, a comprehensive dataset comprising approximately 50,000 surgical cases from various hospitals has been acquired. The paper successfully developed and evaluated various models. The best-performing model was 'Random Forest', which achieved an accuracy of 92.51%. The deep learning models also showed useful results, with the best model achieving an accuracy of 91.3%. The process of fine-tuning the DL model has yielded a significant improvement in accuracy, increasing it from 91.3% to 93.4%. © 2023 IEEE.
30.
Akash, V.; Sai, Monish; Amith, K. B.; Jayanthi, M. G.; Kannadaguli, P.
Lung Nodule Segmentation and Classification Using Conv-Unet Based Deep Learning Proceedings
Institute of Electrical and Electronics Engineers Inc., 2023, ISBN: 979-835030082-6 (ISBN), (0).
@proceedings{98,
title = {Lung Nodule Segmentation and Classification Using Conv-Unet Based Deep Learning},
author = {V. Akash and Monish Sai and K. B. Amith and M. G. Jayanthi and P. Kannadaguli},
doi = {10.1109/NMITCON58196.2023.10276037},
isbn = {979-835030082-6 (ISBN)},
year = {2023},
date = {2023-01-01},
journal = {2023 International Conference on Network, Multimedia and Information Technology, NMITCON 2023},
publisher = {Institute of Electrical and Electronics Engineers Inc.},
abstract = {This paper proposes a Convolutional U-Net architecture, a variation of the standard U-Net architecture for the segmentation of lung nodules and classification using Deep learning on Computerized Tomography (CT) scans. The Primary steps employed are Preprocessing, Segmentation and Classification of nodules. In the preprocessing step, the lung region is segmented using techniques such as normalization, median filtering, Kmeans clustering, morphological and thresholding operations to extract lung Region of Interest (ROI) and nodule masks. The Conv-Unet design adds more convolutional layers to the standard U -Net architecture to help capture complicated patterns and boundaries of lung nodules for more accurate segmentation. Categorization of the segmented lung nodules is done using a CNN network on the LIDC-IDRI, and LUNA16 dataset. Overall, this model achieves a dice score of 62% and classification accuracy of 82% displaying appropriate performance in comparison with other variations of the U-Net architecture. © 2023 IEEE.},
note = {0},
keywords = {AIML},
pubstate = {published},
tppubtype = {proceedings}
}
This paper proposes a Convolutional U-Net architecture, a variation of the standard U-Net architecture for the segmentation of lung nodules and classification using Deep learning on Computerized Tomography (CT) scans. The Primary steps employed are Preprocessing, Segmentation and Classification of nodules. In the preprocessing step, the lung region is segmented using techniques such as normalization, median filtering, Kmeans clustering, morphological and thresholding operations to extract lung Region of Interest (ROI) and nodule masks. The Conv-Unet design adds more convolutional layers to the standard U -Net architecture to help capture complicated patterns and boundaries of lung nodules for more accurate segmentation. Categorization of the segmented lung nodules is done using a CNN network on the LIDC-IDRI, and LUNA16 dataset. Overall, this model achieves a dice score of 62% and classification accuracy of 82% displaying appropriate performance in comparison with other variations of the U-Net architecture. © 2023 IEEE.
31.
Rajani, K. C.; Manjunatha, S.; Rakesh, V. S.; Bhavana, P.
Relational Model and Improvised DSR Assisted Framework for Secure Wireless Communication Proceedings
Springer Science and Business Media Deutschland GmbH, vol. 1861 CCIS, 2023, ISBN: 18650929 (ISSN); 978-303140563-1 (ISBN), (0).
@proceedings{63,
title = {Relational Model and Improvised DSR Assisted Framework for Secure Wireless Communication},
author = {K. C. Rajani and S. Manjunatha and V. S. Rakesh and P. Bhavana},
doi = {10.1007/978-3-031-40564-8_3},
isbn = {18650929 (ISSN); 978-303140563-1 (ISBN)},
year = {2023},
date = {2023-01-01},
journal = {Communications in Computer and Information Science},
volume = {1861 CCIS},
pages = {30-45,},
publisher = {Springer Science and Business Media Deutschland GmbH},
abstract = {Wireless communications offer a wide range of application with cost- effective communication establishment used in both commercial and domestic purpose. Owing to the presence of decentralized environment in Wireless Communications, it is quite a potential issue for monitoring the regular behavior of any wireless node while performing data propagation. The proposed study states that a wireless node in wireless communication environment plays a role of either normal node or malicious node with completely two different polarities of intention. However, the essential concern is mainly towards the selfish node, which is basically a normal node and owing to certain reason their behavior comes under suspicious zone. Irrespective of various forms of security attacks, identifying malicious node behavior is the most challenging security concern in wireless communication, which has yet not received a proper security solution. The proposed work presents a novel idea of security towards resisting adversarial impact of selfish node and routing mis-behavior present in wireless communications environment. The 1st-framework introduces a retaliation model where a novel role of node called as an auxiliary node is introduced to cater up the routing demands along with security demands. This module performs identification of a malicious routing mis-behavior based on evaluation carried out for the control message being propagated to the auxiliary node. The 2nd-framework further introduces a group friendly architecture where a probability-based modelling is carried out considering various possible set of actions to be executed by normal and selfish node. The model also presents a solution towards security threat prevention by encouraging the selfish node to perform cooperation. The study outcome is benchmarked with existing technique to see that proposed system offers a good balance between security performance and data transmission performance in wireless communication. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.},
note = {0},
keywords = {AIML},
pubstate = {published},
tppubtype = {proceedings}
}
Wireless communications offer a wide range of application with cost- effective communication establishment used in both commercial and domestic purpose. Owing to the presence of decentralized environment in Wireless Communications, it is quite a potential issue for monitoring the regular behavior of any wireless node while performing data propagation. The proposed study states that a wireless node in wireless communication environment plays a role of either normal node or malicious node with completely two different polarities of intention. However, the essential concern is mainly towards the selfish node, which is basically a normal node and owing to certain reason their behavior comes under suspicious zone. Irrespective of various forms of security attacks, identifying malicious node behavior is the most challenging security concern in wireless communication, which has yet not received a proper security solution. The proposed work presents a novel idea of security towards resisting adversarial impact of selfish node and routing mis-behavior present in wireless communications environment. The 1st-framework introduces a retaliation model where a novel role of node called as an auxiliary node is introduced to cater up the routing demands along with security demands. This module performs identification of a malicious routing mis-behavior based on evaluation carried out for the control message being propagated to the auxiliary node. The 2nd-framework further introduces a group friendly architecture where a probability-based modelling is carried out considering various possible set of actions to be executed by normal and selfish node. The model also presents a solution towards security threat prevention by encouraging the selfish node to perform cooperation. The study outcome is benchmarked with existing technique to see that proposed system offers a good balance between security performance and data transmission performance in wireless communication. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.
32.
Raj, R. J.; Babu, Anantha; Josephine, V. L.; Varalatchoumy, M.; Kathirvel, C.
Institute of Electrical and Electronics Engineers Inc., 2022, ISBN: 978-166541028-1 (ISBN), (3).
@proceedings{157,
title = {Implementing Multiclass Classification to find the Optimal Machine Learning Model for Forecasting Malicious URLs},
author = {R. J. Raj and Anantha Babu and V. L. Josephine and M. Varalatchoumy and C. Kathirvel},
doi = {10.1109/ICCMC53470.2022.9754005},
isbn = {978-166541028-1 (ISBN)},
year = {2022},
date = {2022-01-01},
journal = {Proceedings - 6th International Conference on Computing Methodologies and Communication, ICCMC 2022},
pages = {1127-1130,},
publisher = {Institute of Electrical and Electronics Engineers Inc.},
abstract = {Web attacks such as spamming, phishing, and malware are common on the Internet. When an unsuspecting user hits the URL, the user becomes a victim of the assaults, which have significant consequences for commercial, finance, and social networking sites. Lexical features, host-based features, content-based features, DNS features, popularity features, and other discriminative features are used to generate a decent feature representation of the URL. URL dataset is collected from ISCX-URL. The goal of this research is to create a multi-class classification model that can categorise URLs as a possible threat to system security by combining several criteria to get the optimal Machine Learning Model. © 2022 IEEE.},
note = {3},
keywords = {AIML},
pubstate = {published},
tppubtype = {proceedings}
}
Web attacks such as spamming, phishing, and malware are common on the Internet. When an unsuspecting user hits the URL, the user becomes a victim of the assaults, which have significant consequences for commercial, finance, and social networking sites. Lexical features, host-based features, content-based features, DNS features, popularity features, and other discriminative features are used to generate a decent feature representation of the URL. URL dataset is collected from ISCX-URL. The goal of this research is to create a multi-class classification model that can categorise URLs as a possible threat to system security by combining several criteria to get the optimal Machine Learning Model. © 2022 IEEE.
33.
Rajani, K. C.; Aishwarya, P.; Manjunath, S.
Friendly Group Architecture for Securely Promoting Selfish Node Cooperation in Wireless Ad-hoc Network Journal Article
In: International Journal of Advanced Computer Science and Applications, vol. 13, pp. 521-527,, 2022, ISBN: 2158107X (ISSN), (1).
@article{117,
title = {Friendly Group Architecture for Securely Promoting Selfish Node Cooperation in Wireless Ad-hoc Network},
author = {K. C. Rajani and P. Aishwarya and S. Manjunath},
doi = {10.14569/IJACSA.2022.0131263},
isbn = {2158107X (ISSN)},
year = {2022},
date = {2022-01-01},
journal = {International Journal of Advanced Computer Science and Applications},
volume = {13},
pages = {521-527,},
publisher = {Science and Information Organization},
abstract = {Wireless Ad-hoc Network is characterized by a decentralized communication scheme with self-configuring nodes which has witnessed a wide range of practical wireless applications. However, this characteristic also results in various security threats in vulnerable wireless environment irrespective of presence of various routing protocols. Review of existing literature shows that there is very less emphasis towards securing Dynamic Source Routing (DSR) while majority of solutions uses encryption-based operation. Therefore, this manuscript introduces a novel non-encryption-based scheme called as Friendly Group Architecture which intends to identify the presence of selfish node followed by presenting a method to promote the secure cooperation of it. The complete modelling is analytically designed using probability-based computation and dynamic thresholding. The simulation outcome carried out in MATLAB exhibits that it outperforms existing system with respect to energy, overhead, and security © 2022, International Journal of Advanced Computer Science and Applications.All Rights Reserved.},
note = {1},
keywords = {AIML},
pubstate = {published},
tppubtype = {article}
}
Wireless Ad-hoc Network is characterized by a decentralized communication scheme with self-configuring nodes which has witnessed a wide range of practical wireless applications. However, this characteristic also results in various security threats in vulnerable wireless environment irrespective of presence of various routing protocols. Review of existing literature shows that there is very less emphasis towards securing Dynamic Source Routing (DSR) while majority of solutions uses encryption-based operation. Therefore, this manuscript introduces a novel non-encryption-based scheme called as Friendly Group Architecture which intends to identify the presence of selfish node followed by presenting a method to promote the secure cooperation of it. The complete modelling is analytically designed using probability-based computation and dynamic thresholding. The simulation outcome carried out in MATLAB exhibits that it outperforms existing system with respect to energy, overhead, and security © 2022, International Journal of Advanced Computer Science and Applications.All Rights Reserved.
