On Information Fusion by Fuzzy Sets for Decision Making in Peer Review

Peer review plays a crucial role in quality insurance of projects, especially for natural science projects, in evaluation or assessment activities. However, the results of assessment by reviewers for a proposal may scatter due to the intrinsic fuzzy attributes in the peer review process. Specifically, it may introduce a review bias with razor-thin margins in the conversion of descriptive opinions into quantified scores. The accumulation of the bias might cause overturning in the results of evaluation, leading proposals toward a twilight zone between approval and rejection. Here, a novel approach to handling scores in evaluation is presented to address the ambiguity of brink in a tight competition, whereby correlation information from multiple sources could be merged to improve the degree of consensus among reviewers for each proposal according to its essential value. This method may provide an alternate evaluation mechanism for proposal reviews and help tackle the challenges in decision intelligence.