JIITA, Vol.10 no.2 pp.1296-1307 (2026), DOI: 10.22664/ISITA.2026.10.2.1296
K. Emayavaramban, Thambusamy Velmurugan
Abstract. the global healthcare landscape, early detection of chronic diseases plays a crucial role in improving patient outcomes and reducing long-term complications. This study focuses on diabetes prediction and risk classification, aiming to identify patients as high, moderate, or low risk based on clinical and demographic features. Traditional machine learning algorithms have been widely used for this task, but they often fall short in prediction accuracy. To address this, we propose a novel ensemble learning method called AXG boosting algorithm, which outperforms conventional algorithms by achieving a 5% improvement in predictive performance and its supports both supervised and unsupervised. The dataset used in this study is sourced from the Kaggle repository and includes key features relevant to diabetes diagnosis. Feature selection techniques are applied to identify the most significant attributes, and the dataset is split into 80% for training and 20% for testing. Experimental results demonstrate the effectiveness of the proposed AXG Boosting algorithm in enhancing risk classification for diabetic patients.
Keywords; Information Boosting algorithm, Machine learning algorithm, AXGB algorithm, diabetes
Fullpaper: