Confident Stock Price Prediction using Markov Chains

The ability to predict future stock price movement is very important to financial institutions that use stock price projections to calculate future values of stocks or indexes. Although there are numerous well-known methods for predicting stock prices, in this paper, we have focused on a short-term Markov chain-based stock price prediction model. Previous studies have demonstrated the practicality of modeling the financial market as a random walk and the potential for perceiving a market to have the Markov property by defining the transition states as deviations from the opening and closing stock prices of the relevant day. The aim of this study was to construct a first-order transition Markov chain-based stock price prediction model, where the states are continuous price ranges as opposed to potential closing price variations. The First-order transition Markov Chain Model was developed on MATLAB 2020. The outcomes include graphs forecasting the future prices for the top five Nifty-Fifty stocks (based on market capitalization at the time) in comparison to the actual stock prices. Further, the confidence interval was generated for the first prediction using a sample size of 100 predictions with respect to the real stock price. The model was recommended for use on a short-term basis as it loses significance over time owing to the volatility of the market and limited transitions found in the used data set.