Enhancing Rural Healthcare with an Automated Ensemble Machine Learning Intelligent Agent for Real-Time Diabetes Diagnosis

This study investigates the efficacy of various machine learning techniques for predicting diabetes, utilizing a dataset sourced from Kaggle that encompasses critical diagnostic features such as glucose levels, BMI, and age. The methodology involved thorough data wrangling and preprocessing steps, including imputation of missing values, outlier handling, feature engineering, and adjustments for class balance. Through exploratory data analysis (EDA), we identified significant indicators of diabetes risk, particularly glucose levels, BMI, and age. We developed and assessed multiple machine learning models, including Decision Tree, Gradient Boosting, Logistic Regression, Random Forest, Support Vector Machine (SVM), K-Nearest Neighbors, and a Voting Classifier, to determine the most accurate predictor of diabetes. Among these models, the Random Forest model exhibited the highest accuracy of 81.7 %, with a commendable F1-Score of 82.4 %, highlighting its superior performance. Furthermore, we implemented an intelligent agent for real-time diabetes diagnosis based on this ensemble machine learning model.