An Improved Hypertension Prediction Model Based on RS and SVM in the Three Gorges Area

There are many unimportant features in the hypertension sample data set in the three gorges area, which are gathered by Tongji medical college, school of HUST. These redundant irrelevant features spoil the classification, increase many unwanted calculations and decrease the real-time capacity of the medical prediction. In order to solve above problem, an improved hypertension prediction model based on rough set and support vector machine is proposed in this paper. Rough set, as an anterior preprocessor of SVM, can find out these relevant factors influencing the hypertension disease by means of greedy attribute reduction algorithm, and then, using them as the input vectors of SVM, the hypertension prediction model is conducted. Experiment results compared with traditional SVM and other machine learning algorithms show that the training rapidity and accuracy of the RS-SVM model are both evidently improved.