Comparative Analysis of Machine Learning Models for Diabetes Prediction

Abstract

This paper focuses on analyzing the benchmark Diabetes dataset which consists of eight commonly measured characteristics. The goal of the study is to present comparative analysis of six machine learning models that predict diabetes, as well as various preprocessing techniques (under-over sampling, feature standardization). The study investigates various approaches and presents results demonstrating that machine learning algorithms can achieve high accuracy results for diabetes prediction, enabling early detection and better outcomes for patients. The paper shows that ensemble learning methods, such as Extra Trees Classifier and Random Forest Classifier, along with appropriate data pre-processing techniques, can lead to 86% accuracy in diabetes prediction classification problems. The paper highlights the potential for machine learning to play a valuable role in the prediction and management of diabetes, leading to improved quality of life and health outcomes for patients.