Prediction of type II diabetes mellitus based on demographic features by the use of machine learning classification algorithms — a study across Assam, India

Abstract

Background

The incidence of type II diabetes mellitus (T2DM) has quadruplicated in the recent decades and Prevention of T2DM cases is possible by changing lifestyle practices. The process of diagnosis of diabetes is a tedious one. The advent and advancement in (AI) techniques presents a probable solution to this critical problem.

Objective

The study aims to assess the diverse attributes of the test sample population across Assam and enhance the early prediction of Type II Diabetes Mellitus by employing artificial neural networks.

Methods

The aim of this study is to design a suitable AI model that prognosticates the likelihood of diabetes in individuals with maximum accuracy based on the levels of liver enzymes. This work also analyzes the effect of fast food intake, sleeping patterns, and consumption of alcohol on healthy controls and contemplates their susceptibility to contract T2DM.

Results

The AI model accurately predicted T2DM likelihood and revealed significant links between unhealthy behaviors and increased T2DM risk among healthy individuals.

Conclusions

The study underscores lifestyle modifications for T2DM prevention, highlighting AI’s potential in diagnosis and the impact of unhealthy habits on T2DM susceptibility.