Modification of triangular member's function (TMF) based on firefly-chen fuzzy time series (FTS) method

Abstract

Usually the applications of fuzzy forecasting methods use triangular fuzzy number for calculating the value of member's function.In many previous researches on forecasting, FTS (FTS) methods were implemented using symmetric triangles in the fuzzification step. In the present paper, we study on how hypotenuse of the triangle can affect forecasting accuracy rates. We modify the TMF to be non-simetric triangle with different line of orientation (left and right leaning triangles) to find out the effect of FTS forecasting results. This modification isapplying on Firefly-Chen FTS method to forecast the Indonesian Composite Stock Price Index (IHSG). Its performance verified through simulation by using Matlab.