Volume - 7 | Issue - 2 | june 2025
Published
22 July, 2025
In recent years, to provide more effective and timely treatment, accurate disease prediction for humans has remained highly challenging. Globally, diabetes is a multidisciplinary disease that threatens people's lives. It mainly targets the organs in the human body, such as the heart, kidneys, nerves, and eyes. By combining healthcare datasets with data fusion techniques and efficient machine learning models, a smart healthcare recommendation system can accurately forecast and suggest diabetes management. There have been recent proposals for machine learning models and methodologies to predict the onset of diabetes. However, these algorithms must remain current when dealing with the diversity of diabetes-related multi-feature datasets. Patients with uncontrolled diabetes are at increased risk of developing diabetes mellitus (DM), a disorder that can endanger many organs. Furthermore, this study delves into the interpretative capabilities of an ML model and the impact of its key components on prediction outcomes. Methods include gathering and organizing large datasets, including demographics, environmental variables, and past medical records. This work also examines the complex interactions with data and identifies patterns that indicate an epidemic of diabetes. The increasing rate of diabetes necessitates effective measures to prevent and manage potential diseases. This research proposes a novel framework for predicting patient diabetes using different machine learning algorithms that utilize Electronic Health Record datasets for experimental purposes. The experimental work, carried out with various evaluation metrics, showed that it outperformed previous methods.
KeywordsMachine Learning Decision Tree Diabetes Mellitus Electronic Health Records Datasets Support Vector Machine
