Identification of Diabetes Disease from Human Blood Using Machine Learning Techniques

Abstract

Diabetes among one of the most common diseases occurs in human beings due to imbalance of insulin level in blood. The early detection of diabetes is very necessary as it can affect many internal parts and immune system of human body silently. In this paper, we are comparing various machine learning and neural network based approaches that are applied on publically available datasets. Here, we have used two datasets for experiments 1st dataset is UCI dataset and other is PIMA Indian dataset then we have performed lots of experiments using different machine learning classifiers and neural network models to observe the performance of each classifier. After experiments, the highest accuracy of identification obtained from decision tree method which is 99.8% for dataset1 and for dataset 2 the highest accuracy was obtained from back propagation neural network model which is 80.8 %.