A Novel Approach to Predict Kidney Detection Using Support Vector Machine

Data Mining is a striking tool for obtaining appreciated information from the huge quantity of available information that can be utilized further for taking the right judgments. Numerous approaches are presented for emerging cost- effective results from the potential data. Mining of data by applying the conditioning rule has been predictably utilized with the objective of the reveling rules in the medical applications. Recognition of different diseases such as kidney, diabetes, and heart attack etc. from huge number of estimates and proofs is an area of great attention for investigators which are not free from false assumptions and unpredictable outcomes. Hence there is terrific requirement to use valuable output resulting from information of patients gathered in our data storehouse. Support Vector Machines, one of the latest approaches for pattern classification, has been widely utilized in large application areas. The objective of this research is to simultaneously optimize parameters and feature subset selection without degrading SVM classification accuracy. We introduce a generic algorithm approach for feature selection named as H-SVD and parameters optimization to solve these kinds of problems. In this paper we design an algorithm that enhances the accuracy of diseases prediction system.