A Modified Weighted Fuzzy Time Series Model for Forecasting Based on Two-Factors Logical Relationship

Abstract

In this paper, we present a modified weighted fuzzy time series model for forecasting based on two-factors fuzzy logical relationship groups. The proposed method define a new technique to partition the universe of discourse into different length of intervals to different factors. Also, the proposed method fuzzifies the historical data sets of the main factor and second factor to their maximum membership grades, obtained by their corresponding triangular fuzzy sets and further constructs the fuzzy logical relationship groups which is based on the two factors to increase in the forecasting accuracy rates. This study also introduces a new defuzzification technique based on the weighted function define on two-factors fuzzy logical relationship groups. The implementation of the proposed method is verified in forecasting on Bombay stock exchange Sensex historical data and compares the forecasted accuracy rate in terms of root mean square and average forecasting error which indicates that the proposed method can achieve more accurate forecasted output over the existing models on fuzzy time series.