Diabetes Risk Prediction Using Machine Learning Methods: A Case Study of Logistic Regression, Random Forest, and Support Vector Machine

Abstract

Diabetes is a global health problem whose prevalence continues to increase and has the potential to cause serious complications if not detected early. Early detection of diabetes risk is very important to prevent fatal consequences and support more effective clinical decision-making. This study aims to analyze and compare the performance of three machine learning algorithms, namely Logistic Regression, Random Forest, and Support Vector Machine, in predicting diabetes risk based on patient symptom and health condition data. The methods used include data collection from open sources, data pre-processing with imputation and normalization, model training using the three algorithms, and model performance evaluation based on accuracy, precision, recall, and F1-score metrics. The dataset used consisted of 520 samples with 17 attributes, including main symptoms such as frequent urination, excessive thirst, and demographic factors. The results of the study show that the Random Forest algorithm has the best performance with an accuracy of 99%, precision of 1, recall of 0.99, and F1-score of 0.99, surpassing Logistic Regression and Support Vector Machine. These findings confirm that Random Forest is very effective in capturing the complexity of clinical data and can be relied upon as a tool for early detection of diabetes risk. This research contributes to the development of decision support systems in the field of digital health by offering an accurate and efficient machine learning-based prediction approach.