Diabetes Prediction with Random Forest - Logistic Regression algorithm using voting classifier

A commonly recognized chronic metabolic condition known as Diabetes mellitus significantly affects the global, social and economic standing of people. Obesity, age, high blood pressure, glucose levels, pregnancy, heredity, and ethnicity are some of the variables that might lead to diabetes mellitus. Estimates indicate that by 2030, 578 million people worldwide will have diabetes, more than double the number recorded in 2010. Several approaches has been used so far to predict the diabetes. The usage of machine learning algorithms in clinical practices has expedited the recognition of individuals with diabetes for medical professionals. Furthermore, to more precisely anticipate a person’s diabetes, supervised machine learning is employed. In order to predict diabetes, the Random Forest-Logistic Regression technique is presented in this paper by using a classifier called soft voting classifier. From the results, it is observed that Random Forest-Logistic Regression approach improves prediction accuracy by 1.75% in comparison to the effective Support Vector Machine (SVM) technique.