Performance Analysis of Data Mining Classification Algorithms to Predict Diabetes

Data mining refers to non-trivial extraction of valid, implicit, novel, potentially useful and ultimately understandable information patterns of data from enormous volumes of data. Classification and prediction are two forms of data analysis that can be used to extract models describing important data classes or to predict future data trends. One of the most important application of data mining is in disease prediction. In this paper we present a classification model developed using cloud platform Microsoft Azure that predicts the occurrence of Diabetes in an individual on the basis of non-pathological parameters – age, gender, family history of being diabetic, smoking and drinking habits, frequency of thirst and urination, weight height and fatigue. Six different algorithms have been compared among which the model created using “Two-Class Neural Network Algorithm” has the highest accuracy of 98.3% and hence has been deployed as a web service. Finally, a GUI is been developed in python to access the web service.