Stock Market Price Behavior Prediction Using Markov Models: A Bioinformatics Approach

Summary

Financial stock market price behavior prediction is very important problem which can be expressed briefly using computational biology or bioinformatics techniques. Markov Models have newly used in stock market price behavior prediction. In this work, we have applied the zero order, first order, second order, and Hidden Markov Model (HMM) techniques for forecasting the daily stock market price behavior (i.e. increasing and non‐increasing) in form of 1's and 0's for given closing price of day to day financial data. The fluctuations are characterized by a number 1and 0, where ‘0' denoting non‐increasing behavior (i.e. Sensex either steady or falling down) and ‘1' denoting increasing behavior (i.e. Sensex rising day to day) in the financial Stock market. The behavior of financial market 0 and 1 are changes to nucleotides A, T, C, and G. Now behavior of financial market converted in to biological DNA sequence and we have used the zero order, first order, second order, and Hidden Markov Model (HMM) procedure to predict future behavior of daily stock market Sensex. In this work, our motivations is prediction of stock market behavior through bioinformatics tools like HMM and get successfully closest behavior results from the historical training Sensex data to the test sequences. Finally we have drawn a comparison analysis between all four models. In future other higher order Markov Models have explored.