Presenting a model for predicting the Tehran Stock Exchange Index using ANFIS and fuzzy regression

Abstract

The purpose of this study is to provide a prediction model for the Tehran Stock Exchange Index using Adaptive Neuro-Fuzzy Inference System (ANFIS) and fuzzy regression analysis. The behavior of this index is nonlinear and chaotic that traditional methods do not predict accurately. Hence, using the above two tools and identifying three macroeconomic variables including inflation rate, exchange rate and crude oil price as independent variables, we predicted the index of the total stock index for the next week. Then, the modeling was performed using the above three variables. By comparing the results, ANFIS performance was better than fuzzy regression. The Root Mean Square Error Performance criterion was obtained for the ANFIS output of 0.021248. The prediction of the next week showed an error reduction for both tools and ANFIS again with an error value of 0.007933, yielded superior performance of the study. Also, the model with four inputs was more accurate compared to the model with three inputs. The emphasis on using macroeconomic variables, predicting the next week's index number, using the two tools mentioned, analyzing the sensitivity of the models during the research are the characteristics of this research. This research can be used by all companies in the stock exchange, investors, brokers, and individuals and legal entities dealing in any way with the stock market.