Title: An improvement of hidden Markov model for stock market predictions
Authors: Seyed Kazem Chavoshi; Azadeh Mansouri; Sorour Sheidani
Addresses: Department of Business, Faculty of Management, Kharazmi University, Tehran, Iran ' Department of Engineering, Faculty of Computer Engineering, Kharazmi University, Tehran, Iran ' Department of Business, Faculty of Management, Kharazmi University, Tehran, Iran
Abstract: This paper predicts Tehran Exchange Dividend and Price Index (TEDPIX) by finding a pattern in TEDPIX through settled transactions and open orders volume effects. To do so, we improve an autoregressive hidden Markov model (AR-HMM) by adding a more hidden layer. Then, we utilised a genetic algorithm for long term daily trend predictions. By exploiting the obtained information of predicted five days using the genetic algorithm, we update the parameters of improved AR-HMM. This stepwise prediction-updating process continues until all desired number of future days stock exchange indices get predicted. Comparing our new scheme with other studied Markov family models shows that the added features lead to achieve more accuracy and less prediction errors. Experimental results show that mean absolute percentage error of all predictions by our improved AR-HMM approach are less than 5% which indicates far better performance of our method against Markov and Hidden Markov Models.
Keywords: autoregressive; hidden Markov models; TEDPIX; settled transactions; open orders.
International Journal of Reasoning-based Intelligent Systems, 2022 Vol.14 No.2/3, pp.144 - 153
Accepted: 02 Feb 2022
Published online: 09 Sep 2022 *