AI AND DATA ANALYTICS FOR PROACTIVE HEALTHCARE RISK MANAGEMENT

Abstract

This work aims at examining the use of AI in anticipation of diabetes and the use of data analysis. To achieve this, we trained different classifiers using a broad Diabetes Detection Dataset which includes the following; Logistic Regression, K Nearest Neighbor (KNN), Random Forest Classifier, and Decision Tree Classifier. The observation made from the analysis showed that the Random Forest Classifier yielded the highest overal accuracy of 96. It was found that 82% of the population was affected, with a precision of 0.94 and a recall of 0.69 for positive cases. The overall performance of KNN model was also impressive it had a precision of 0.91 and a recall of 0.62, while LR and DTC gave useful information about the data but did not perform so well in some of the evaluation measurements. Some of these items included correlation heatmaps and ROC curves that helped in capturing the relationship between diabetes and other health aspects such as blood glucose level, HbA1c and model performance. It confirms that the concept enshrined in the AItechnologies actually has a huge potential in the early diagnosis and intervention programs that will eventually lead to efficiency enhancement and harmonization of health care services delivery. Future research should hence focus on issues to do with the generalizations of the models as well as ways of combining data in order to enhance the level of predictive and health care improvement. This study predicts diabetes with the aid of machine learning models. Exploratory data analysis shows that the major predictors of diabetes are age, blood glucose, and hypertension. The implementation of the models Logistic Regression, KNN, Random Forest, and Decision Tree was done. Random Forest turned out to be the best model being very accurate and precise in predicting negative cases and quite reasonable in performance for positive case detection. This research contributes to the early detection of diabetes and potentially better treatment for patients.