Training and Classification of PCA with LRM model for Diabetes Prediction

Abstract

There are exponential increase in the number of families who are diagnosed by diabetes mellitus because of lifestyle and other non-determinable factors. Most of the patients are least bothered about the consequences they face or about the danger factor that approaches them. In this, we have established a novel model predicting the type 2 diabetes mellitus (TD2M) dependent on information digging methods. The main constraints are that we are trying to enhance the precision of the expected model and to not limit the model with just one data set. The model contains the improved NB, DT, KSTAR, LOGISTIC REGRESSION, SVM compared to the pre-processing techniques. To compare our outcome and the outcomes from different scientists we use Pima Indians diabetes data set and the Waikato environment for knowledge analysis toolbox. Apart from these, the model which we expect to implement have adequate data set quality. For more analysis, we applied it to two more diabetic datasets. These two provides satisfied outcomes. Henceforth, the model is set to be valuable for the betterment in the field of diabetology..