The Relationships of Prior Information and Interval Partition on the Forecasting Effect of Fuzzy Time Series Two-Factor Model

Abstract

This study uses two data sets, Taiwanese export values as the prediction varialbe and its foreign exchange spot rates as the auxiliary variable, to discuss two forecasting issues in the fuzzy time series analysis by using One-and Two-factor models. First, disregarding the length of information period, when window basis is two, a better forecasting result is obtained. Second, the optimum number of partition equal intervals is to select 14 intervals to have the smallest MSE in all models. However, if partitioning the information into more than 14 equal intervals, MSE can not be reduced but presents a waving pattern.