Suicide Prediction Using Machine Learning on Assamese Language Manabjyoti Choudhury, Aniruddha Deka, Gunikhan Sonowal, Soraisam Gobinkumar Singh Proceedings of the 4th IEEE International Conference on Interdisciplinary Approaches in Technology and Management for Social Innovation Iatmsi 2026, 2026 Sentiment analysis in low-resource Indian languages remains a challenging research area due to limited annotated corpora, complex linguistic morphology and script-specific processing requirements. This study presents a vigorous machine learning channel for Assamese sentiment arrangement using a stratified TF-IDF feature representation combined with a stacking ensemble classifier. The proposed framework integrates Random Forest, XGBoost, Gradient Boosting and Decision Tree as base learners with meta-level soft voting to enhance generalization. Experiments were conducted on a manually curated three-class Assamese sentiment dataset and the model performance of 87.58 % of accuracy and 94.45 % of macro-AUC value in ensemble stacking. The projected method accomplishes good performance compared to existing deep learning and transformer-based models reported in recent Assamese and Indic sentiment literature, demonstrating that optimized ensemble learning remains a strong contender for resource-constrained languages where pretrained language models or large-scale datasets may not be accessible.
OPTIMIZED CNN-BASED APPROACH FOR ALZHEIMER’S DISEASE BY TACKLING CLASS IMBALANCE IN MRI CLASSIFICATION Soraisam Gobinkumar Singh, Dulumani Das, Barman, Utpal, Huseynov, Hasan Reliability Theory and Applications, 2025 Accurate and early diagnosis of Alzheimer’s Disease (AD) is crucial for effective intervention and treatment. This study presents a Convolutional Neural Network (CNN)-based approach for the classification of brain MRI images into four categories: Mild Demented, Moderate Demented, Non-Demented, and Very Mild Demented. To address the challenges of class imbalance inherent in the dataset, we employed class weighting and focal loss during training. Class weighting ensured that underrepresented classes received adequate attention, while focal loss emphasized harder-to-classify examples, resulting in improved model performance on minority classes. The model achieved remarkable results, with an accuracy of 97.66%, precision of 97.66%, recall of 97.66%, F1-score of 97.66%, specificity of 98.98%, and Cohen's Kappa of 96.14%, indicating a robust performance across all metrics. A comparative analysis with state-of-the-art methods demonstrated that our approach outperformed many existing models, including Siamese CNNs, 3D DenseNet ensembles, and other transfer-learning-based techniques. The ROC-AUC analysis further highlighted the model's ability to distinguish between classes with near-perfect curves for all categories. These results underscore the effectiveness of combining CNN architectures with class imbalance-handling strategies for medical image classification. The proposed method holds promise for improving diagnostic accuracy and early detection in AD, thereby supporting clinical decision-making.
COMPREHENSIVE EVALUATION AND PERFORMANCE ANALYSIS OF A DEEP LEARNING MODEL WITH HYPERPARAMETER TUNING FOR LUMPY SKIN DISEASE CLASSIFICATION IN DAIRY COWS Gunikhan Sonowal, Soraisam Gobinkumar Singh, Bairagi, Prasanta, Barman, Utpal, Dulumani Das, et al. Reliability Theory and Applications, 2025 This work attempts to classify lumpy skin conditions using CNN and hyperparameter tuning. This model is comprised of many procedures, including selecting a pre-trained model, altering the architecture, and training the model on a specific dataset. During tweaking, the proposed model attained a validation accuracy of 89.73 percent. The model’s generalisation performance was confirmed with an accuracy of 80.68% in the final test set evaluation. It significantly increased the timeliness of LSD identification, making it a valuable tool for farmers and veterinarians. Furthermore, a Receiver Operating Characteristic (ROC) curve with an Area Under the Curve (AUC) of 0.88 indicates that our binary classifier performed satisfactorily.
Early Alzheimer’s Disease Detection: A Review of Machine Learning Techniques for Forecasting Transition from Mild Cognitive Impairment Soraisam Gobinkumar Singh, Dulumani Das, Utpal Barman, Manob Jyoti Saikia Diagnostics, 2024 Alzheimer’s disease is a weakening neurodegenerative condition with profound cognitive implications, making early and accurate detection crucial for effective treatment. In recent years, machine learning, particularly deep learning, has shown significant promise in detecting mild cognitive impairment to Alzheimer’s disease conversion. This review synthesizes research on machine learning approaches for predicting conversion from mild cognitive impairment to Alzheimer’s disease dementia using magnetic resonance imaging, positron emission tomography, and other biomarkers. Various techniques used in literature such as machine learning, deep learning, and transfer learning were examined in this study. Additionally, data modalities and feature extraction methods analyzed by different researchers are discussed. This review provides a comprehensive overview of the current state of research in Alzheimer’s disease detection and highlights future research directions.
RECENT SCHOLAR PUBLICATIONS
Suicide Prediction Using Machine Learning on Assamese Language M Choudhury, A Deka, G Sonowal, SG Singh 2026 IEEE International Conference on Interdisciplinary Approaches in … , 2026 2026
Integrating XAI with Optimized CNNs: A Novel Approach for Imbalanced Alzheimer's Disease SG Singh, D Das, U Barman Proceedings of Fifth Emerging Trends and Technologies on Intelligent Systems … , 2025 2025
OPTIMIZED CNN-BASED APPROACH FOR ALZHEIMER’S DISEASE BY TACKLING CLASS IMBALANCE IN MRI CLASSIFICATION SG Singh, D Das, U Barman, H Huseynov Reliability: Theory & Applications 20 (SI 7 (83)), 197-207 , 2025 2025
COMPREHENSIVE EVALUATION AND PERFORMANCE ANALYSIS OF A DEEP LEARNING MODEL WITH HYPERPARAMETER TUNING FOR LUMPY SKIN DISEASE CLASSIFICATION IN DAIRY COWS G Sonowal, SG Singh, P Bairagi, U Barman, D Das, M Iltimas, ... Reliability: Theory & Applications 20 (SI 7 (83)), 58-65 , 2025 2025
Early Alzheimer’s disease detection: a review of machine learning techniques for forecasting transition from mild cognitive impairment SG Singh, D Das, U Barman, MJ Saikia Diagnostics 14 (16), 1759 , 2024 2024 Citations: 45
MOST CITED SCHOLAR PUBLICATIONS
Early Alzheimer’s disease detection: a review of machine learning techniques for forecasting transition from mild cognitive impairment SG Singh, D Das, U Barman, MJ Saikia Diagnostics 14 (16), 1759 , 2024 2024 Citations: 45
Suicide Prediction Using Machine Learning on Assamese Language M Choudhury, A Deka, G Sonowal, SG Singh 2026 IEEE International Conference on Interdisciplinary Approaches in … , 2026 2026
Integrating XAI with Optimized CNNs: A Novel Approach for Imbalanced Alzheimer's Disease SG Singh, D Das, U Barman Proceedings of Fifth Emerging Trends and Technologies on Intelligent Systems … , 2025 2025
OPTIMIZED CNN-BASED APPROACH FOR ALZHEIMER’S DISEASE BY TACKLING CLASS IMBALANCE IN MRI CLASSIFICATION SG Singh, D Das, U Barman, H Huseynov Reliability: Theory & Applications 20 (SI 7 (83)), 197-207 , 2025 2025
COMPREHENSIVE EVALUATION AND PERFORMANCE ANALYSIS OF A DEEP LEARNING MODEL WITH HYPERPARAMETER TUNING FOR LUMPY SKIN DISEASE CLASSIFICATION IN DAIRY COWS G Sonowal, SG Singh, P Bairagi, U Barman, D Das, M Iltimas, ... Reliability: Theory & Applications 20 (SI 7 (83)), 58-65 , 2025 2025
