Analysis of Diabetes mellitus using Machine Learning Techniques
Salliah Shafi Bhat, Venkatesan Selvam, Gufran Ahmad Ansari, Mohd Dilshad Ansari
Abstract
Diabetes Mellitus (DM) often known as hyperglycemia is caused by high blood sugar levels. Although DM is a metabolic chronic disease. Treatment and early detection are essential to reducing the risk of serious outcomes. The World Health Organization (WHO) reports that diabetes has a significant mortality rate causing 1.5 million deaths worldwide. The disease can be identified early because to technology tremendous improvements. In order to build a model with a few variables based on the PIMA dataset this research focuses on evaluating diabetes patients as well as diabetes diagnosis using various Machine Learning Techniques (MLT). Exploratory data analysis is the first step in our process after which the information is transferred for data pre-processing and feature selection. The relevant features are chosen and the data is then training and testing using three different MLT such as Support Vector Machine (SVC), Random Forest (RF) and K-Nearest Neighbors (KNN). Amongst all of the classifiers Random Forest has the highest accuracy of 97.75% followed by Support Vector Machine (82.25%) and K-Nearest Neighbors (86.25%).