Insurance Credit Evaluation Incorporated Qualitative and Quantitative Information

The insurance credit evaluation often involves qualitative and quantitative information. This paper presents a unified framework to synthesize and aggregate them. In this framework, the qualitative information is expressed by the linguistic labels or expressions, and the quantitative information is expressed by the numerical values. The semantics of linguistic expression is described by a semantic similarity relation defined on the finite set of linguistic labels. By transforming the linguistic expression into a semantic similarity distribution on the set of linguistic labels, a semantic score can be obtained consequently. Then the transformed semantic scores from the qualitative evaluations are aggregated with the quantitative information using the weighted average operation. This novel method solves the semantic gap between qualitative description and quantitative description in a natural way, and has potential application in risk evaluation and credit evaluation of financial market.