Early Detection and Prediction of Diabetes Using Ensemble Classifier
Yash Prajapati, Darshan G Hihoriya, Shanti Verma
Abstract
Diabetes is a chronic disease that affects millions of people worldwide. It is a complex condition that can be caused by a variety of factors, including pregnancy, blood pressure, glucose levels, and BMI. Identifying individuals who are at risk of developing diabetes is crucial in the prevention and management of this disease. According to a World Health Organization (WHO) report, in India 8.7% of the population suffers from diabetes in the 20 to 70 years. India is also the second most affected country in the world from this disease. Early detection and prediction from this disease is necessary to help citizens of the county to reduce the adverse conditions. In this paper authors try to build an ensemble classification model based on demographics, pregnancy, blood pressure, glucose levels, and BMI using bagging, boosting and averaging methods. . Authors used a secondary dataset available on kaggle.com. The results of the study says that classification models built using random forest algorithms have higher accuracy than other algorithms which is 81.1%. The accuracy of the model is not satisfactory so authors applied ensemble learning methods averaging, bagging and boosting and found the averaging method has less error than other methods. The results of the study are helpful in the healthcare industry for early prediction and detection of this disease.