Diabetes Prediction using Logistic Regression and Support Vector Machine (SVM) Classifier

Early detection and diagnosis of diabetes are crucial for managing and treating the disease effectively. Diabetes is a chronic condition characterized by the body’s inability to produce sufficient insulin or use it properly, resulting in high blood sugar levels. This can lead to serious health issues, such as heart disease, nerve damage, strokes, kidney failure, and vision problems. Identifying diabetes early can help individuals live longer and healthier lives by managing the condition promptly. This study aims to predict diabetes using various machine learning algorithms and compare their performance. We utilized a dataset containing numeric attributes related to patients’ health indicators. The data has been preprocessed and analyzed using Exploratory Data Analysis (EDA) and two machine learning algorithms were implemented: Logistic Regression and Support Vector Machine (SVM). The models are evaluated based on accuracy, precision, and recall, F1-score, and ROC curve. Our results indicate that the SVM algorithm outperformed the Logistic Regression model with the highest accuracy of $\mathbf{9 8 \%}$ and a ROC-AUC score of $\mathbf{0. 9 8}$. This study highlights the potential of machine learning in the early detection of diabetes and provides a comparative analysis of different classification algorithms. The findings suggest that SVM can be an effective tool for diabetes prediction, potentially aiding healthcare professionals in making informed decisions. Future work could focus on incorporating additional features and exploring advanced ensemble methods to further improve prediction accuracy.