Group aggregation method of non-formatted text evaluation information based on linguistic label and evidence distance

A group decision-making model for non-formatted text evaluation information is suggested. First, the problem is transformed into a particular multi-criterion decision making frame based on multi-granular linguistic label, which includes the comprehensive evaluation score and the evaluation criterion point score. Second, a weighted model for the criteria is proposed based on the minimal difference between the comprehensive evaluation score and the evaluation criterion point score for decision makers (DMs). Third, a numerical estimation model for the incomplete evaluation item is established based on the minimal evidence distance of Dempster-Shafer theory, which is maximal group consistency. In addition, a weight calculation method for the DM is suggested through the logistic consistency of the DM and the similarity among the groups. Furthermore, the score modification method for the alternative is proposed according to the weight of the criteria and the DM preferences. Finally, a case of a soft science project evaluation and selection illustrates the application process and the feasibility of the proposed method.