Stochastic Multiple Criteria Comprehensive Evaluation Based on Probabilistic Linguistic Preference Relations: A Case Study of Healthcare Insurance Audits in China

Abstract

As a typical multiple criteria comprehensive evaluation (MCCE) problem, healthcare insurance audits (HIA) play a vital role in governmental audits. To handle the HIA in an autonomous region in China, a systematic methodology in the framework of MCCE is presented. An evaluation system is constructed at first based on the specific auditing objective. The group of individual preferences is collected by a probabilistic linguistic preference relation (PLPR). Its consistency and prioritization are handled based on the idea of stochastic analysis and the widely acknowledged techniques of linguistic preference relations (LPRs). Specifically, the consistency degree is measured by the probability of stochastically generated LPRs being with acceptable consistency, and then improved by revising the probabilistic distributions involved in the PLPR. The priority is derived by the expected value of priorities of the stochastically generated LPRs. Moreover, the comprehensive evaluation is conducted based on aspiration-based utility functions. The proposed methodology enables decision makers to determine parameters intuitively, and presents interpretable and admissible results. It is effective even if the original individual preferences are with low consistency degree. Finally, the focused HIA problem is solved based on the proposed methodology.