Classification of heart failure with Polynomial Smooth Support Vector Machine
Yu-Bo Yuan, Wen-Qiang Qiu, Yingjie Wang, Ju Gao, Ping He · 2017
With the development of machine learning techniques, artificial intelligence applications in medicine are becoming hot topic in health information systems. In this research, we construct a new basic heart failure disease database which contains 1715 patients and 400 features. Then, we propose a new machine learning method called Polynomial Smooth Support Vector Machine(PSSVM) to help doctors diagnose heart disease, after solved by BFGS method, the algorithm parameters are obtained and they can be used to determine the state of patients' heart disease. In order to solve the problem of high dimension of database and enhance the speed of PSSVM, we use PCA, L-DA, CCA and LPP methods to decrease the features. Finally, we compare the performance of our method with LibSVM and ELM. We empirically demonstrate the effectiveness of our approach by comparing its performance with LibSVM and ELM.