A fuzzy preference-based Dempster-Shafer evidence theory for decision fusion

Dempster-Shafer evidence theory (D-S) is an effective instrument for merging the collected pieces of basic probability assignment (BPA), and it exhibits superiority in achieving robustness soft computing decision making uncertain imprecise environment. However, determination BPA still uncertain, merely applying can sometimes lead to counterintuitive results when lines conflict. In this paper, a novel generation method binary problems called as base algorithm designed based on kernel density estimation construct function models, using pairwise learning establish classification pairs. By means new method, decision-making D-S theory, fuzzy preference relation nondominance criterion effectively designed. The strength proposed presented learning, which transforms original complex problem into simple subproblems. With process, complexity be solved greatly reduced, increases feasibility industrial applications. Furthermore, technique used aggregate output each single subproblem, degree class determined from matrix, directly input instance. Based several industrial-based experiments, present effectiveness improvement terms precision Cohen’s kappa.