Abstract
A new linguistic granular model is proposed and the effect of its parameters on the output is analyzed. The design of the model consists of two stages: using conditional fuzzy clustering for information granular, and integrating all information granules to final output. The integrating tool is fuzzy integral based on fuzzy measure, and the generalization of fuzzy integral increases flexibility of the linguistic granular model greatly. A heuristic algorithm to determine the parameters in the fuzzy integral is used to realize the linguistic model. Two experiments verify the feasibility of the proposed model.