Analysing Machine Learning Techniques in Python for the Prediction of Diabetes Using the Risk Factors as Parameters

Abstract

Diabetes is a disease caused due to an increased level of glucose in the blood. Many people around the world suffer from this condition. A lot of research works have been published concerning diabetes, using a variety of classifiers and machine learning models, both for predictions and diagnosis. This paper outlines the effects of some risk factors of diabetes, specifically the effect of BMI, blood pressure and physical activity, using machine learning models. A survey was conducted among a random sample of Indians of various age groups and the data of 251 individuals was gathered after which ML algorithms were applied. The analysis is supported by naïve Bayes, logistic regression and random forest algorithms. All the metrics of these algorithms are compared and discussed in this paper, thereby establishing the model having the highest accuracy for the prediction. The model with the highest accuracy and precision can be used or analysed for further computational research on diabetes mellitus.