The Prediction Of Stock Market Based On Support Vector Machine

Stock market is a complex non-linear system, and is affected by many factors. The traditional prediction technologies cannot disclose the inherent rule of stock market. In this paper, a new prediction technology based on Support Vector Machine (SVM) has been proposed. SVM can be used to solve many problems that traditional technologies cannot solve effectively.First, this paper introduce the background knowledge of stock market, then traditional prediction technologies are introduced in detail, especially the technology based on neural network, and then the basic principles of SVM are discussed. Second, this paper uses SVM to predict the price of stock, and propose a common framework to solve stock market prediction problems using SVM. Data from real stock market is used to evaluate the exactness of the algorithm. Result shows that SVM is an effective method, and get precise result.Third, In order to improve the efficiency of the algorithm, this paper presents major factor extraction method to optimize the input vector of SVM. Result shows that this method can get similar result while using less computation time and less storage consumption.At last, this paper researched the problem of kernel function selection, and compared four types of kernel functions. Suitable kernel functions are chosen. This research will improve the precision of prediction. The research shows SVM is very suitable to solve this problem. It is believed that SVM will be an important method in the problem of predicting stock market.