Diabetes mellitus detection using machine learning techniques

Abstract

Diabetes Mellitus (DM) is a common disease that is spreading worldwide and affecting millions of people. The use of machine learning techniques for the timely and accurate diagnosis of diabetes is the subject of this study. This research makes use of a dataset that includes measures like insulin resistance metrics and blood glucose levels in addition to clinical and demographic data. The important factor leading to the growth of diabetes mellitus is the lifestyle which includes lack of exercise, anxiety, age factor, family history etc. It is a serious condition, if not cured, can lead to several health problems like heart disease, nerve damage, kidney damage. A range of machine learning algorithms and deep learning techniques such as support vector machines, decision trees, and neural networks, are utilized to examine and simulate the connections present in the data. Techniques for feature extraction and selection are applied to improve the models' performance and highlight important factors that lead to the development of diabetes. The goal of the study is to attain high sensitivity and specificity to guarantee the accurate identification of Type 1 and Type 2 diabetes.