Similarity measure based on nonlinear compensatory model and fuzzy logic inference

In this paper, we propose a novel nonlinear nearest-neighbor (NNN) matching for similarity measure based on nonlinear compensatory (NC) choice model. Based on fuzzy logic inference, we propose NC choice model which granulates the psychological boundary between linear and nonlinear compensatory in the decision-making. Based on our NC mode, we develop a NNN matching function to consider both linear and nonlinear psychological compensatory effects. Theory analysis and experiment have demonstrated the success of NNN matching and NC model.