A Fuzzy Qualitative Evaluation System: A multi-granular aggregation approach using fuzzy compound linguistic variable

Abstract

Measuring qualitative attributes is a complex process which includes imprecise decisions for designing, rating, and quantifying the qualitative attributes of the measurement objects. This paper proposes a Fuzzy Qualitative Evaluation System FQES using expert judgment to evaluate perceived qualitative attributes. FQES highlights the fusion of the capabilities of humans and computers in the qualitative evaluation process as humans are proficient in fuzzy reasoning and classification, while computers are superior in calculating human fuzzy inputs with embedded fuzzy algorithms derived in this paper. The innovative methods for the fusion of the capabilities includes a Fuzzy Compound Linguistic Variable FCLV for two dimensional rating categories, a Double Step Rating Approach for determination using FCLV, a Parabola-based Fuzzy Normal Distribution FND for assigning fuzzy numbers for the FCLV and a 5-layer Fuzzy Analytical Network 5FAN which is the multi-granularity aggregation system taking multiple fuzzy number inputs to infer the result using an parametric function. FQES could be the ideal framework to increase the assessment accuracy for qualitative evaluation applications such as customer satisfaction surveys, employee satisfaction surveys, vendor surveys and risk assessment surveys. In the application, this paper demonstrates how FQES can be used to identify the optimum number for the selection of suppliers.