Study On The Prediction Of Shanghai Composite Index Based On GA-Markov Chain Model

With the development of stock market in recent years, the stock investment has become an important mode of capital management. The investors focus on how to analyze and predict the trend of stock market accurately to improve the gains. Nowadays with the rise of artificial intelligence technology, many domestic and foreign researchers have studied and explored how to apply the artificial intelligence technology into stock price prediction. This paper presented a new approach which was the combination prediction model of genetic algorithm and Markov chain to optimize the radial basis function neural network and the support vector regression respectively in order to predict the stock closing price of the Shanghai composite index accurately.This paper completed the work, as follows.① Explorer current status of studies on the stock market prediction from domestic and overseas and illustrate the basic knowledge of related technology for the discussions below;② Established two single models which were the radial basis function neural network model and support vector regression model to forecast the closing price of Shanghai composite index. They employed the same data as training sample and testing sample;③ Adopt the Markov chain optimized by genetic algorithm to correct the relative error between actual value of the stock closing price and predicted value based on radial basis function neural network model and support vector regression model respectively. The experimental results indicated that the prediction accuracy of the hybrid model which was the radial basis function neural network model optimized by the combination of genetic algorithm and Markov chain was higher than the radial basis function neural network model and radial basis function neural network model optimized by the Markov chain. And the prediction accuracy of the hybrid model which was the support vector regression model optimized by the combination of genetic algorithm and Markov chain was higher comparing with the support vector regression model and support vector regression model optimized by the Markov chain. In addition, this paper compared these two hybrid models and found that the accuracy of the support vector regression model optimized by the combination of genetic algorithm and Markov chain was higher than the radial basis function neural network model optimized by the combination of genetic algorithm and Markov chain.