Classification of Diabetes by Kernel Based SVM with PSO

Abstract

Background: The classification method is required to deduce possible errors and assist the doctors. These methods are used to take suitable decisions in real world applications. It is well known that classification is an efficient, effective and broadly utilized strategy in several applications such as medical diagnosis, etc. The prime objective of this research paper is to achieve an efficient and effective classification method for Diabetes. Methods: The proposed methodology comprises two phases: The first phase deals with t h e description of Pima Indian Diabetes Dataset and Localized Diabetes Dataset, whereas in the second phase, the dataset has been processed through two different approaches. Results: The first approach entails classification through Polynomial Kernel, RBF Kernel, Sigmoid Function Kernel and Linear Kernel SVM on Pima Indian Diabetes Dataset and Localized Diabetes Dataset. In the second approach, PSO has been utilized as a feature reduction method followed by using the same set of classification methods used in the first approach. PSO_Linear Kernel SVM provides the highest accuracy and ROC for both the above mentioned dataset. Conclusion: The present work consists of a comparative analysis of outcomes w.r.t. performance assessment has been done PSO and without PSO for the same set of classification methods. Finally, it has been concluded that PSO selects the relevant features, reduces the expense and computation time while improving the ROC and accuracy. The used methodology could be implemented in other medical diseases.