Crime data optimization using neutrosophic logic based game theory

Abstract

SummaryThe significance of prisoner's dilemma in making two completely rational individuals might not cooperate even their best interest to do so. Every criminal investigation is strongly reliant on crime data classification and optimization. In this work, we extended prisoners' dilemma for crime data classification and its optimization. A confusion matrix‐based optimization technique for crime data with game theory model in order to identify the person involvement in crime is to be predicted and clustered them into confess, not confessed, and indeterministic states based on the neutrosophic logic principle. In this technique, first we map neutrosophic logic into crime system to break the uncertainties in crime clustering. Next, we preliminary cluster the crime data pairs that indicates two offenders neutrosophic values into three clusters. However, some pairs lie in intersections of two or three clusters. So we optimize the clustering into disjoint clusters will be done based on the ratio of intra‐cluster and inter‐cluster distances. We have implemented the proposed method. The experimental result shows that the quality of proposed method based on accuracy, precision, and recall parameters, observed more than 90% accuracy.