Machine Learning for Smartphone-Based Early Detection of Diabetic Disease in Pima Indians Diabetes Database

Abstract

Now a day’s Diabetic disease(Dd) is common disease among people which causes damage to kidney, heart and may eventually lead to death. Early detection of diabetics is very much essential to avoid kidney and heart failure. Effective treatment for Dd are available through it requires early diagnosis and the continuous monitoring of diabetic patients. Also many physical tests can be used to detect Dd but are time consuming. The objective of our research paper is to give decision about the presence of diabetics by applying ensemble of machine learning classifying algorithms on features extracted from output of different datasets. It will give us accuracy of which algorithm will be suitable and more accurate for prediction of the disease. Decision making for predicting the presence of diabetic is performed using Linear  Regression(LR), Logistic Regression(LoR), K Nearest Neighbors(KNN), Decision Tree(DT), Support Vector Machine(SVM), Naïve Bayes(NB), Random Forest(RF), The experimental results has been analysed by using Jupyter Notebook. Among all the mentioned Supervised machine learning algorithm RF approach show a highest classification accuracy (CA) of 89.58. From this, we can infer that for diabetic the RF approach gives the best performance compared to all other approaches.