Research Multi-source Evidence-based Analysis Of Evidence Theory

With the dramatic changing development of information technology, networks andelectronic information have emerged in people’s daily life, they gave rise to a huge convenienceto obtain and process information, they meanwhile bring the risks and challenges. Manyinformation security issues exposed such as fraud using the network information, crime inducedby network micro-blog, and crime through mainframe computers. The traditional approaches likethe scene and papery forensics can’t completely solve the problems of this kind, the cases can’tbe solved as soon as possible, then we need effective tools and means to make analysis about theelectronics forensic. During the procedure for the analysis of the evidence, we will inevitablyencounter situations like too much information on the amount of data, the evidence informationcome from different areas, different evidence sources of information, and the evidence orinformation with uncertainty is difficult to define and judge. Evidence theory is a commonlyused inference method in recent years due to their own advantages and it is widely used toprocess uncertain information. It can merge multiple sources of evidence information effectively,and can narrow the number of assumptions space collection. Evidence theory has a goodperformance in decision-making problems and information fusion problems in many areas.Evidence theory is doing the analysis after the combination and mergence of the evidence, butgiven the large amount of evidence to be analyzed, too complex evidence information anddifferent sources of information, the paper analyzes the fuzzy clustering algorithm and improvesit for pretreatment of evidence, discusses the problems that evidence theory exists and suggestsimprovements for combination rules. Then the improved fuzzy clustering algorithm and evidencetheory are applied to the analysis process for multiple sources of evidence. The ultimate basicprocess is evidence collection, pretreatment for evidence, information fusion for multiple sourcesof evidence, evidence analysis. The main points of work this article does is as follows:(1) Discuss the basic concepts of evidence theory and fuzzy clustering algorithmsystematically, elaborate the combination rules, advantages, problems and solutions of evidencetheory and introduces theoretical basics for research of fuzzy clustering algorithm requires.(2) Study and learn fuzzy C-means clustering algorithm which is a widely applied algorithmin fuzzy clustering algorithms. Analyze and describe the basic meaning of the algorithm, thebasic processes and the problems of this algorithm. Introduce the simulated annealing algorithmto determine the initial number of clusters for fuzzy C-means clustering algorithm to avoid thetraditional algorithm limitations based on a priori knowledge to determine this figure, andmeanwhile the clustering center function is weighted so that the speed of the improved functionconverges to the desired target accelerated and the efficiency is greatly improved.(3) Research and learn the Evidence theory deeply, probe into the basic concepts,composition rules, advantages and problems of this theory. In this paper, it gives a method toimprove the combination rule of evidence theory for the problems caused by exponential growthof information fusion of the focal elements. Besides, to improve the combination rule andnumerical experiments combination rules, to verify the improved information fusion method andreduce the amount of calculation. (4) Gather the process evidence information from the host forensics. Preprocess theevidence information using the improved fuzzy C-means clustering algorithm, then fuse theprocessed multisource information using the improved evidence theory, finally make an analysisand decision of the evidence. Form a multisource evidence analysis technique based on theevidence theory to analyze multiple sources of evidence effectively.