Predicting Diabetes Using Outlier Detection and Multilayer Perceptron with Optimal Stochastic Gradient Descent

Now a day's diabetes is becoming a major health issue in the world for men and women. To extract meaningful data from medical datasets or databases we can use techniques of data mining and machine learning models. In this article, for classifying the data properly machine learning model called multilayer perceptron (MLP) with optimal stochastic gradient descent (SGD) proposed. Radial basis function (RBF) model as an outlier detection method is used for removing the misclassified instances then removed instances passed to the multilayer perceptron with a stochastic gradient descent classifier model for classifying the data more effectively. Performance of MLP-SGD is good to compare to other classifiers but the usage of RBF took the performance of the proposed model to the higher level compared to existing models.