Telematique

Severity Aware Early Diabetes Prediction Using Hybrid Machine Learning and Deep Learning Techniques

K H Rizwana, Dr. Ajay Sharma — Wed, 10 Dec 2025 00:00:00 +0000

In the recent trend and in medical domain, diabetes stands out as one of the most predominant disease and unfortunately considering as a life-threatening disease. Early prediction of diabetes plays a pivotal role in initiating prompt treatment and halting the progression of the condition. The proposed methodology not only aids in predicting the future diabetes but also finds its severity scores. By presenting this issue as a multi-class classification problem, hybrid machine learning (ML) and deep learning (DL) techniques are used to build the new hybrid model. This helps in incorporating both structural feature learning of ML and deep temporal pattern recognition of DL for better performance. The hybrid ML+DL for diabetes prediction used XGBoost, LightGBM, CatBoost ML models and Temporal Convolutional Network (TCN) as base layer, Logistic Regression (LR) as a meta-classifier. The model is evaluated and fine-tuned for effective diabetes disease prediction with its score of severity. The experimental findings underscore the effectiveness of each component in the framework and its impact on the accuracy. The proposed work proves the sufficient amount of accuracy as 99.4%, and HML+DL compared with the recent studies in prediction of early stage of diabetes.

Mapping Research Themes and Future Directions in Learning Style Detection Research: A Bibliometric and Content Analysis

Ms. Neethu D S, Dr. Ajay Sharma — Mon, 12 Jan 2026 00:00:00 +0000

This study aims to provide a comprehensive overview of the current state and potential future research in learning style detection. With the increasing number and diversity of research in this area, a quantitative approach is necessary to map out current themes and identify potential areas for future research. To achieve this goal, a bibliometric and content analysis will be conducted to map out the existing research and identify emerging topics and directions for future research. The study analyzes 1074 bibliographic sources from Scopus and visualizes the results of the bibliometric analysis through cooccurrence and thematic map analysis using VOSviewer and BibliometriX software. Content analysis is then conducted based on the results of the co-occurrence analysis. The findings reveal a significant increase in publications and citations in the field, with popular research topics including classification, adaptive learning, and MOOCs, and the most frequently used learning style models being Felder-Silverman, VARK, and Kolb. Emerging research topics include the use of EEG signals, online learning, and feature extraction. Future research may focus on classification, intelligent tutoring systems, MOOCs, online learning, adaptive learning, and deep learning. This study provides valuable insights into the current and future research trends in learning style detection, which can support the development of adaptive e-learning systems, intelligent tutoring systems, and MOOCs. By identifying popular research topics and emerging areas of study, this research can guide the design and implementation of effective online learning environments. Additionally, the study advances the field of e-learning knowledge by providing a comprehensive overview of the most frequently used learning style models and potential research areas. It sheds light on the ongoing development of learning style detection research and the potential for future advancements in the field, ultimately contributing to the growth and improvement of e-learning practices.

A study of Radiological Image-Based Bone sarcoma Detection Using Transfer Learning

P. J. Adit, Dr. C. Priya — Thu, 15 Jan 2026 00:00:00 +0000

Bone sarcoma occurs primarily in children, adolescents and adults. Diagnostic assessment has traditionally involved subjective and often time-consuming assessments of imaging modalities such as X-ray, MRI and CT scans. This paper introduces a framework based on deep learning for automated classification of Bone Sarcoma from standard imaging modalities. The modernization utilizes MobileNetV2, a labeled dataset trained on ImageNet, to efficiently extract significant features while limiting computation. Preprocessing of the dataset included image normalization, resizing and augmentation using flipping, rotation, changing the zoom level and adding contrast. The dataset had a split of 80% affected and remaining 20% unaffected, respectively. During the fine-tuning phase the last layers of the model were unfrozen, and models trained at a reduced learning rate to accommodate the imaging data specific to Bone Cancer. The model trained with Adam optimizer and binary cross-entropy loss function with about 93% training accuracy and over 90% validation accuracy. Using evaluations from precision, recall, F1, and confusion matrix, the results verified the model robustness with minimal false negative rates being crucial for medical diagnostic. The results indicated that the suggested approach also provides a reliable, lightweight, and accurate diagnostic support for radiologists.