Stock market forecasting model based on AR(1) with adjusted triangular fuzzy number using standard deviation approach for ASEAN countries

Abstract

Traditional autoregressive (AR) time series models have been extensively applied to predict various stationary data sets based on single point data. However, real-world system involves uncertainty due to human behaviours and incomplete information. Since the single point data is not able to represent the nature of data, fuzzy approach is necessary to deal with such uncertainties in the analysis. This paper proposes AR(1) model building based on triangular fuzzy numbers. A procedural step for building triangular fuzzy number based on standard deviation approach is provided, to handle the existence of uncertain information and the biasness during data collection. The proposed model is applied to forecast buying–selling stock market prices by using real data sets from five ASEAN countries. The results from this study show that the proposed method with triangular fuzzy numbers exhibits smaller error. That is, the proposed method is able to achieve almost similar accuracy performance as obtained by the traditional autoregressive approach, yet it also solves the uncertainties issue in the analysis.