The Prediction of Stock Price Movements. Testing the Predictive Power

Stock market, especially prediction stock market movement, has been always attracted people’s attention from professional to scientific field of activity. This is perfectly justified due to the fact that development an ideal stock price prediction model can bring ultimate insights to the investor or trader how they should act in the future. Unfortunately, stock price prediction is complicated task not only for the top-class trader, but also for the scientific researcher because stock market cannot be purely described by the mathematical models alone. According to the financial literature, it is not possible to outperform the market with sophisticated predictive algorithms or filter rules due to the efficiency of the stock market. But with the development of computational ability of the machines people challenge the Efficient market hypothesis and try to earn abnormal positive return in comparison with broad market. The main goal of this paper is to develop perfect prediction model based whether on LSTM or Markov switching model and apply it for the trading algorithm in order to check the ability to outperform the market in term of brining significant returns.