Research Of Evidence Fusion Method Based On DBSCAN Clustering

With the rapid development of computer network technology and the emergence of information physics fusion system,information fusion technology has received more attention.How to quickly,accurately,and comprehensively analyze complex and diverse information has become a research hotspot.Therefore,studying an effective mechanism to solve the problem of information fusion is very meaningful.Evidence theory has become the basic method of information fusion because of its strong inferential ability.In order to dealing with the paradox of high-conflict information,the paper proposes an evidence fusion method based on DBSCAN((Density-Based Spatial Clustering of Applications with Noise)density clustering,and we design a multi-source information fusion system model.The main achievements of the research are as follows:(1)A comprehensive conflict measurement method is proposed.As we known,solving the problem about seriously conflict information is mainly based on the comprehensive measurement and reasonable allocation of the conflicts.After focusing on the conflict measurement method,the paper proposes a comprehensive conflict metric method combined Jousselme's evidence distance and focal element sequence distance.On the one hand,the Jousselme distance is used to reflect all the differences between the focal elements of the evidence.On the other hand,the overall output decision of the evidence is considered from the order of the size of the evidence focal elements,so that the conflict can be comprehensively characterized and compared with the existing conflict metrics.The method is more comprehensive and provides an effective distance function model for subsequent density clustering analysis.(2)Based on DBSCAN clustering,an evidence fusion method is proposed.In the mechanism,the clustering between evidences is taken into consideration and the density clustering algorithm is introduced to mine potential information.The method firstly uses DBSCAN density clustering to preprocess the source evidence set.In the cluster analysis model,the above CCMF is used to correctly measure the conflict evidence;then,based on the DBSCAN density clustering results,two different methods are used respectively.The discount factor is used to modify the original evidence set;finally,the modified evidence set is synthesized using the DS combination rules.(3)Results and analysis.Based on DBSCAN clustering and D-S evidence theory in this paper,a multi-source information fusion system model is proposed.Besides,Simulation results show that the proposed algorithm has better performance about validity and stability by comparing with other researchers' fusion methods.