Stock Index Forecasting by Hidden Markov Models with Trends Recognition

Stock index forecast is a complicated problem since the financial market is influenced by various underlying factors and some of them are unobservable. Hidden Markov Model (HMM) is an effective probabilistic graphic model which can models the hidden states of the observed sequence. So the underlying patterns of the stock movements can be learnt by HMM models. In this paper, HMM is employed to forecast the Shanghai Stock Exchange Composite Index and Shenzhen Stock Exchange Composite Index, which are the two representative stock indices in China. The proposed method is empirically tested on the target data sets and compared with other models based on HMM using Mean Absolute Percentage Error (MAPE). Results show that the proposed HMM model achieves good forecasting performance, and at the same time reveals meaningful hidden states of stock market.

I. Introduction

Stock market dynamics always associate with national financial market change, as well as various individual investment activities. It not only reflects the movement of virtual economics, but also suggests modifications of substantial industries and thus influence the macroeconomy to some extent. Especially for the extreme case of financial crisis, the stock market may reveal early warning signals and also could result in an extreme high volatility status. Stock index is a sort of financial indicator which is calculated by a group of representative stock prices. It comprehensively characterize the level and the change of market prices. Stock index forecasting may serve as an early warning mechanism to monitor the associated financial market, and provide meaningful insights for investment as well as government regulations. One of the most acknowledged challenge of stock index forecasting is that there are so many influencing factors and the behind mechanisms are mostly either unknown or uncertain. Data driven models are always criticised of lacking theoretical foundations, yet most ideal financial propositions fail to show acceptable performance in forecasting. There is always a call for theoretically sound models with high forecasting performances.