Predicting missing pairwise preferences from similarity features in group decision making

In group decision-making (GDM), fuzzy preference relations (FPRs) refer to pairwise preferences in the form of a matrix. Within field GDM, problem estimating missing values is utmost importance, since many experts provide incomplete preferences. this paper, we propose new method called entropy-based for FPR. We compared accuracy our algorithm predicting with best candidate from state art achievements. proposed method, took advantage achieve good results by storing extra information single rating scores, example, comparison alternatives vs. alternative’s score one five stars. The maps prediction into matrix factorization problem, and thus solution can be expressed latent expert features alternative features. Thus, embeds same feature space. By virtue embedding, another novelty approach use similarity experts, as well between alternatives, infer even when only minimal data are available some experts. Note that current approaches may fail any output such cases. Apart values, salient contribution paper rank alternatives. It worth mentioning ranking have possible applications especially recommendation systems (GRS).