Font Size: 

Various rule-based fuzzy inference systems (FISs) have been widely used for time series forecasting. Because the Type-1 fuzzy regression functions (FRFs), a kind of FISs, instead of using a rule-based approach, determine the relations between the inputs and outputs by utilizing a series of linear regression models, it does not have this problem. However, due to the nature of the FRFs' inputs, the multicollinearity problem is an inevitable issue that leads to unstable estimates and often gives misleading information. This study presents a liu-type fuzzy regression functions (Liu-FRFs) approach to deal with collinearity problem in time series forecasting.The shrinkage and biasing parameters of Liu-FRFs are determined through the genetic algorithm (GA).The outstanding forecasting performance of the proposed Liu-FRFs is proven via different implementations.