A Modified SVM Classifier Based on RS in Medical Disease Prediction

Too many unimportant attributes are ended up specifying in medical disease sample data sets if we are not sure which attribute to include for disease prediction, which could spoil the classification and increase many unwanted calculations of the medical disease prediction. Thus how to preprocess these medical data and enhance the prediction performance is worth a problem to research. In the paper, a modified SVM classifier based on RS is proposed in medical disease prediction. RS not only provides new scientific logic and research method for information and cognitive science, but also develops effective preprocessing techniques for intelligent information process. It can find out these relevant features influencing the medical disease. And then, using them as the input vectors of SVM, the medical disease prediction model is conducted, which make great use of the advantages of RS in eliminating redundant information and take full advantage of SVM to train and test the data. Experiment results explain the validity and feasibility of our proposed algorithm.