ABSTRACT
Diabetes is a vicious illness with no cure, and there has been an increase in patient rates over the last decade. The disease is induced when the beta cells in the body cannot produce adequate insulin. In this chapter, both boosting and bagging methodologies have been used to deplete overfitting and bias. This chapter also depicts the results obtained by a voting classifier. The Pima Indian Diabetes dataset (courtesy of Kaggle) was used, which contains data from 768 people with different attributes such as insulin level, body mass index, and blood pressure, which all are symptoms ascertaining diabetes. Thus, the machine learning algorithm could be trained using this dataset, making it meticulous with each phase. The highest accuracy among these algorithms was achieved using random forest, 79% under the standard.
