Enhancing the Prediction of Diabetics using Bagging Ensambler Classifier

Abstract

Diabetes is a chronic disease exemplified by far above the ground level of blood sugar. It can up shot in a variety of complex diseases such as a stroke, a kidney failure, a heart attack, and so on. Widely, in the year of 2021 survey report based on diabetes affected almost nearly 512 million humans. The primary aspiration of this research investigation is to develop a machine learning (ML)-based system for predicting diabetic patients. In this paper, the disease of diabetics is to predict using “Machine Learning Classification and ensemble techniques” on a PIMA diabetes dataset. The Machine learning classifier techniques similar to “Logistic Regression”, “Naive-Bayes”, “KNN”, “Support Vector Machine”, “Decision Tree”, “Bagging meta-estimator” and “Random Forest classifier” are used for prediction. The proposed method employs various classification and ensemble methods that are implemented using Python. There were numerous missing values in the dataset; it will be resolved using the median replacement method. The performance factors of “Classification Accuracy, Precision, Recall and F1-Score” are measured using these machine learning classification methods. Finally, based on the results, we concluded that the Bagging Ensemble Method has the highest accuracy compared to other techniques.