STOCK PRICE FORECASTING BASED ON BACK PROPAGATION NEURAL NETWORK AND MARKOV CHAIN by Azme Bin Khamis & Li Chee Guan , Scientific Research Journal (Scirj), Volume VIII, Issue VII, July 2020 Edition

Scirj Volume VIII, Issue VII, July 2020 Edition
ISSN: 2201-2796

STOCK PRICE FORECASTING BASED ON BACK PROPAGATION NEURAL NETWORK AND MARKOV CHAIN

Azme Bin Khamis & Li Chee Guan

Abstract: This study focused on the stock price prediction of an oil and gas exploration company in Malaysia. An artificial neural network that trained by back propagation algorithm or back propagation neural network (BPNN) as well as a hybrid model that combining back propagation neural network and Markov chain are used in this study. The main objective of this study is to compare the forecasting performance of the hybrid model and the BPNN model. The sample data is taken from 1st October 2018 to 30th September 2019. The main interest in this study is daily closing price, and the attributes are opening price, highest price, lowest price, trading volume, Brent crude oil price and OPEC oil price. The hybrid model is built by initialized the forecasting using back propagation neural network, followed by Markov state division, transition probability matrix formulation and prediction computation. Accordingly, both predictive models are compared by Mean Absolute Percentage Error (MAPE) and coefficient of determination R2. As a result, back propagation neural network is better than hybrid model because it showed a higher level of forecast accuracy with 0.5057% of MAPE and 95.94% of R2.

Reference this Paper: STOCK PRICE FORECASTING BASED ON BACK PROPAGATION NEURAL NETWORK AND MARKOV CHAIN by Azme Bin Khamis & Li Chee Guan published at: "Scientific Research Journal (Scirj), Volume VIII, Issue VII, July 2020 Edition, Page 57-61 ".

Search Terms: Artificial Neural Network, Markov chain, Back Propagation Algorithm, Forecast Accuracy

[Read Research Paper] [Full Screen]