Stock Market Prediction Based On Support Vector Machine

The stock market is a complex system which is full of different kinds of information, so it’s difficult to predict rises and falls of stock market. Usually investors wants to predict rises and falls with the usage of different information, in order to get excess returns.Support vector machine(SVM) is a classification theory derived from computational learning theory. It is now widely used in computational learning and pattern recognition fields thanks to its outstanding characters in classification, regression etc. SVM provides a new method to predict rises and falls of stock market.This paper combines the domestic and foreign stock markets, widely exercises SVM to market index, sector index and individual share. This paper uses MATLAB to build the SVM model, tests 4 composition indices,5 sector indices and 9 individual shares, and simulates the net-value changes using SVM investment strategies, achieves reasonable results and luckily gets ideal results in Chinese non-ferrous metal market. We can conclude SVM model is a useful choice for investment.