M-factors Fuzzy Time Series for Forecasting Stock Price Movement

Abstract

Forecasting stock price movements is one of the vital activities in assisting to analyze stock market movements technically. However, the forecasting process is inspiring to many researchers in attempting to apply and modify the forecasting methods due to the difficulties and uncertainties of the stock market.One of the continuity innovation forecasting methods in this area is fuzzy time series (FTS). This paper proposed M-factor FTS using weighted subsethood in constructing the FTS model (WeSuFTS model). 1-factor WeSuFTS and 2-factors WeSuFTS models were developed based on WeSuFTS forecasting procedure. Modeling and evaluation part of the data analysis were obtained from the one company listed in Bursa Malaysia website. The daily historical data were selected for two months. The results from step by step algorithm were demonstrated in this paper.Investigation on the accuracy of forecasting results were compared between actual value and M-factor FTS forecast value in evaluation part of the data using absolute percentage error (APE), mean square error (MSE), mean absolute percentage error (MAPE) and root mean squared error (RMSE).  The models performance results revealedthe proposed M-factor FTS model can further be improve due to 1-factor WeSuFTS model is the outperform model compared to 2-factors WeSuFTS model to forecast stock price movement with forecasting error from 6.3% - 13.8% (APE), 8.8% (MAPE), 0.0041 (MSE) and 0.0638 (RMSE).