Prediction Of Stock Price Index With Hidden Markov Model

Three problems can be solved with Hidden Markov Model (HMM), they are matching, learning and decoding. In this thesis, we use HMM to develop a system to predict the stock price index. Specifically, we make Gauss distribution as our observation sequence density function. After training, we obtain a local optima, and the S&P500 indexes are chosen as our sample set.HMM was known for the first time because of its application in Speech Recognition. Now it has been used for a great many fields. As for stock and other financial time series predicting, Anderas S. Weigend and Shanming Shi put forward the Hidden Markov Expert (HME), HMM was used for prediction of financial time series by Yingjian Zhang, Md. Rafiul Hassan and Baikunth Nath imposed HMM to stock price forecasting. Experimental results show that our model can beat S&P500 Index well.