Title: Prediction of stock price based on hidden Markov model and nearest neighbour algorithm

Authors: Asadullah Al Galib; Mahbub Alam; Rashedur M. Rahman

Addresses: Department of Computer Science, University of Calgary, 2500 University Drive, NW, Calgary, Alberta, T2N1N4, Canada ' Department of Electrical Engineering and Computer Science, Plot-15, Block-B, Bashundhara, Dhaka 1229, Bangladesh ' Department of Electrical Engineering and Computer Science, Plot-15, Block-B, Bashundhara, Dhaka 1229, Bangladesh

Abstract: Stock prices change every day as a result of market forces and other economic factors. The price fluctuation results into unpredictable supply (buy) and demand (sell) of the shares over a span of time. In this research, we have used hidden Markov model on the input stock data and parameters to create a composite model which can predict an important entity of the stock market, the next day's opening price. The precision and the truthfulness of the output is then compared with another model constructed by the nearest neighbour algorithm, whose main foundation lies behind the fact that stock event/data reflects its own behaviour along the time span. The result found in this research is encouraging and offers a new paradigm for stock market forecasting.

Keywords: stock data; Dhaka stock exchange; DSE; hidden Markov model; HMM; nearest neighbour; Bayesian model; opening price; financial time series modelling; price patterns; Markov process; non-parametric class of regression; stock price prediction; stock prices; market forecasting.

DOI: 10.1504/IJIDS.2014.064446

International Journal of Information and Decision Sciences, 2014 Vol.6 No.3, pp.262 - 292

Published online: 10 Sep 2014 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article