Computer Science, Information Systems and Management
25
Scopus Publications
Scopus Publications
Modified oversampling based borderline SMOTE with noise reduction techniques for IoT smart farming dataset M. Suresh, S. Manju Priya International Journal of System of Systems Engineering, 2026 Global population is anticipated to grow exponentially to 10 billion in the future years. To feed the globe, agriculture must be prioritised. Agriculture is vital to human survival. Every field plant breeding, agricultural monitoring, automated maintenance systems, sensor use, and agrochemicals has evolved physiologically and technologically. Technology and analytics merge in Internet of Things (IoT)-based farm data. Machine learning algorithms analyse massive agricultural data. Predictive analytics learning algorithms built with machine learning are fast and effective. The data pipeline's pre-processing stage uses the SMOTE with noise reduction, an advanced oversampling technique. This unique pre-processing method is rigorously compared to SMOTE, ADASYN, and NRAS to assess its efficacy and robustness. This comparison analysis evaluates our improved method's precision, recall, and accuracy in class imbalance scenarios, a common machine learning challenge. To increase synthetic sample quality and model prediction, address dataset noise and borderline occurrences. This research claims that pre-processing affects machine learning models, especially with skewed data. Dataset preprocessing and WSVM performance analysis. The precision, sensitivity, f1-score, accuracy, specificity, and time consumption of MSBNRT are superior.
Adaptive Fuzzy C-Means with logit boost distributed clustering for cancer detection with protein sequences K. Thenmozhi, M. Pyingkodi, V. S. Prakash, Kripa Josten, S. Manju Priya, J. Vennila Discover Applied Sciences, 2025 The Adaptive Fuzzy C-Means with Logit Boost Distributed Clustering (AFC-LBDC) technique is introduced to enhance cancer detection promptly. The various conventional techniques often struggle to improve cancer detection due to their high complexity effectively. In contrast, the AFC-LBDC technique groups similar protein sequences to get better accuracy in cancer detection. Initially, a large protein dataset is divided into ‘C’ number of local clusters using an adaptive Fuzzy C-Means distributed clustering approach. For any protein sequences that are not assigned to a group, the Bayesian probability is computed to find the higher chance of the protein sequence becoming a member of a specific cluster. The Logit Boost technique is applied to improve the clustering performance further, which combines the number of local clusters to make a global cluster. The proposed AFC-LBDC method demonstrates high accuracy rates of 96%, 88%, and 86% for the P53, BRCA2, and HRAS cancer datasets, respectively. Comparative evaluation reveals that AFC-LBDC reduces cancer detection time by up to 31% compared to existing methods, achieving a 20% and 31% reduction over the RaNC and IDMPhyChm-Ens methods for the P53 dataset, 19% and 31% for BRCA2, and 22% and 32% for HRAS. Likewise, the proposed method significantly lowers the false positive rate, with reductions of 27% and 39% for P53, 28% and 36% for BRCA2, and 23% and 31% for HRAS, compared to RaNC and IDMPhyChm-Ens, respectively. In addition, AFC-LBDC minimises space complexity by up to 44%, with 27% and 39% reductions for P53, 24% and 42% for BRCA2, and 22% and 44% for HRAS datasets. These results collectively indicate the superior performance and efficiency of AFC-LBDC in cancer gene detection. The global clustering result improves the cancer detection accuracy and minimises the false positive rate.
IoT-based intelligent infrastructure decision support system with correlation filter and wrapper framework for smart farming M. Suresh, S. Manju Priya International Journal of Critical Infrastructures, 2025 Agriculture is the backbone of the Indian economy in a world where the market is battleground, and technology is constantly changing. More than 75% of the population relies on this ancient craft. Each farmer must produce high-quality harvests despite water shortages and plant illnesses. They must delicately balance soil nutrients, sustaining fertility like a nation's lifeline. From these trials emerged the modern Indian farmer's hero: an IoT-based decision support system, a smart agricultural beacon. This miracle anticipates agricultural yield and guards their livelihood like a sentinel. It monitors soil fertility, stops soil degradation, and considers excessive irrigation a crime against nature. Wireless sensor devices elegantly communicate data to a central server to arrange this technology symphony. In the digital world, a machine learning system does predictive irrigation. The weather, soil, rainfall, seed damage, drought, and alchemical pesticides and fertilisers are considered. Many pioneers in this growing industry have failed, resulting in incorrect estimates and low crop yields. CBF-SF, an artisanal hybrid correlation-based filter (CBF) and sequential forward wrapper architecture is the solution. This clever technique turns parched areas into bountiful goldmines by predicting crop yields with precision, making farmers contemporary alchemists.
Scalable Congestion Management in Ultra-Dense IoT Environments using Adaptive Control J. Gokulapriya, P. Logeswari, V. Yuvaraj, S. Manju Priya, Prema K, Ragunath K Proceedings of the 9th International Conference on Electronics Communication and Aerospace Technology Iceca 2025, 2025 Unparalleled expansion of ultra-dense IoT networks enabled by 5G and next-generation technologies is threatening to cause severe congestion control, mobility, and resource allocation challenges, particularly to delay-sensitive applications. In this work, there is a proposed scalable and delay-sensitive adaptive congestion control protocol (SD-ACCP), the main innovation of which is the two-feedback adaptive mechanism that integrates local real-time feedback, distributed congestion signaling, adaptive transmission control, and priority-oriented packet scheduling. The SD-ACCP facilitates the detection of congestion in a system and allocation of priorities to delay-sensitive traffic and modulates the rate of transmission of each packet on the fly to reduce the number of packets lost and delay, without interfering with the low-priority flows. Lightweight congestion signatures also provide energy-efficient transmission and scalability of ultra-dense networks. It has been demonstrated that SD-ACCP works better than baseline protocols with a packet delivery ratio of 96.2, end to end delay of 85 ms and saving a lot of energy, achieving a network lifetime of up to 4200 seconds. These results demonstrate that SD-ACCP could provide a novel, trustworthy, low-latency, energy efficient method of scaling and sustainable management of congestion in dense IoT systems.
Empowering Farmers with AI: Risk Mitigation Through Machine Learning in Smart Irrigation Systems Sumi M, S Manju Priya Proceedings of International Conference on Circuit Power and Computing Technologies Iccpct 2024, 2024 The advent of IoT-enabled smart irrigation systems in the domain of precision agriculture has revolutionized farming practices, significantly enhancing crop yield and resource management. However, this technological transformation has exposed farmers to novel challenges, particularly in the realms of security, privacy, and risk management. This research investigates the vital utilization of machine learning (ML) for intrusion detection in the case of smart farming, particularly in utilizing IoT technology to enhance and automate irrigation processes is a hallmark of modern precision agriculture, with the aim of mitigating risks. Farmers today face an array of risks, from crop-specific threats such as pests and diseases to broader environmental factors like weather fluctuations. The effective prediction of these risks is essential for sustainable agriculture. Employing ML techniques, this research endeavors to develop an advanced intrusion detection system that not only safeguards smart irrigation infrastructure but also predicts and mitigates risks associated with crop health and water resource management. This holistic approach leverages data from various IoT sensors, including soil-moisture sensors, weather stations, and drone-based imagery, to provide immediate response into potential risks. While benefits in this technology are immense, it also brings security and privacy challenges, as the data generated and transmitted by these systems can be susceptible to unauthorized access and malicious manipulation. Our important aims are to equip farmers with an integrated solution that not only detects and prevents intrusions into their smart irrigation systems but also provides predictive insights into potential agricultural risks.
Innovative Approaches to Agricultural Risk with Machine Learning Sumi. M, S. Manju Priya International Journal of Advanced Computer Science and Applications, 2024 —Agriculture is fraught with uncertainties arising from factors like weather volatility, pest outbreaks, market fluctuations, and technological advancements, posing significant challenges to farmers. By gaining insights into these risks, farmers can enhance decision-making, adopt proactive measures, and optimize resource allocation to minimize negative impacts and maximize productivity. The research introduces an innovative approach to risk prediction, highlighting its pivotal role in improving agricultural practices. Through meticulous analysis and optimization of a farmer dataset, employing pre-processing techniques, the study ensures the reliability of predictive models built on high-quality data. Utilizing Variation Inflation Factor (VIF) for feature selection, the study identifies influential features critical for accurate risk classification. Employing techniques like KNN, Random Forest, logistic regression, SVM, Ridge classifier, Gradient Boosting and XGBoost, the study achieves promising results. Among them KNN, random forest, Gradient Boosting and XGBoost scored with high accuracy of 88.46%. This underscores the effectiveness of the proposed methodology in providing actionable insights into potential risks faced by farmers, enabling informed decision-making and risk mitigation strategies.
A Robust Cardiovascular Disease Predictor Based on Genetic Feature Selection and Ensemble Learning Classification Sadiyamole P. A., S. Manju Priya International Journal of Electrical and Computer Engineering Systems, 2023 Timely detection of heart diseases is crucial for treating cardiac patients prior to the occurrence of any fatality. Automated early detection of these diseases is a necessity in areas where specialized doctors are limited. Deep learning methods provided with a decent set of heart disease data can be used to achieve this. This article proposes a robust heart disease prediction strategy using genetic algorithms and ensemble deep learning techniques. The efficiency of genetic algorithms is utilized to select more significant features from a high-dimensional dataset, combined with deep learning techniques such as Adaptive Neuro-Fuzzy Inference System (ANFIS), Multi-Layer Perceptron (MLP), and Radial Basis Function (RBF), to achieve the goal. The boosting algorithm, Logit Boost, is made use of as a meta-learning classifier for predicting heart disease. The Cleveland heart disease dataset found in the UCI repository yields an overall accuracy of 99.66%, which is higher than many of the most efficient approaches now in existence.
A Hybrid Feature Selection Technique to Predict Breast Cancer 14th International Conference on Advances in Computing Control and Telecommunication Technologies Act 2023, 2023
A NOVEL APPROACH BASED ON VOTING ENSEMBLE AND PCA DIMENSIONALITY REDUCTION METHOD FOR THE PREDICTION OF HEART DISEASE Journal of Theoretical and Applied Information Technology, 2022
Ensemble of multi objective optimizer with pareto frontier solutions for feature selection in large- scale microarray rule datasets Journal of Green Engineering, 2020
Modified whale optimization algorithm for feature selection in micro array cancer dataset International Journal of Scientific and Technology Research, 2020
Complexity analysis of compressing genomic sequence data with chained hash indexing in multiple dictionary-based LZW International Journal of Recent Technology and Engineering, 2019
Improved malaria prediction using ant colony feature selection and random forest tree in big data analytics Journal of Advanced Research in Dynamical and Control Systems, 2018
The implementation of hybrid MFSVM in cross sectional views of brain MRI segmentation in the diagnosis of ADHD Journal of Advanced Research in Dynamical and Control Systems, 2018
Improved ant colony on feature selection and Weighted Ensemble to Neural Network Based multimodel disease risk prediction (WENN-MDRP) classifier for disease prediction over big data International Journal of Engineering and Technology Uae, 2018
