A Method of Evidence Association Mining on Target Identification
Xiaoxuan Wang, Hongjie Wang · 2017
In the target identification system, it is difficult to get the target identification result because of massive and heterogeneous data obtaining from various information sensors. In this paper, a probability statistical method is used to analyze the quantification method through the association analysis of support evidences and the fact of the evidence. Using this novel algorithm, a large number of evidences which are not related to the facts could be filtered out quickly, and the evidences of different aspects which reflect the comprehensive identification of targets could be extracted. Meanwhile, the evidence history information classification model and the evidence-related intensity matrix are established, and the model is updated iteratively according to the new evidences contained in multiple information sources in the target identification system. Based on this method, this paper carries out a simulation experiment of evidence correlation of target identification, which verifies the feasibility and validity of the method.