Research On Information Fusion Based On Rough Set And Evidence Thoery

Rough set theory is an approach of data analysis dealing with imprecise, inconsistent and incomplete information, on which the data analysis and processing technology based are applied more widely. Evidence theory is a powerful tool handling the uncertain problem. As a common method of information fusion, it is significant to solve the uncertainty in the fusion. Based on the relationships between rough set and evidence theory, the thesis focuses on information fusion and conflict information integration processing.The thesis observes the concepts and properties of rough set and evidence theory, analyzes the corresponding relationships between the two theories and utilizes the classification and dependability of rough set to decompose the complex decision. Then it optimizes the decision sub-table based on attribute significance and further changes the decision attribute, to compute the basic credibility. According to the rules of evidence theory, the idea of evidence synthetic method based on the rough theory is given as to achieve the synthesis of evidence.As for research on the integration of conflict evidence, the thesis states the emergence and solution of conflict evidence, and detailed the experts' opinions about its synthetic methods and classified comparison. Moreover, in view of the combination of conflict evidence, the paper provides a combination method of conflict evidence based on a weight. According to the relations among the focal elements and among the evidences, the paper determines the value of evidence, introduces average evidence and modifies the evidence model. The approach can deal effectively with a high degree of conflict, raise the capacity of handling uncertain problems and also avoid conclusion inconsistent with reality.Through simulation, the effect and accuracy of conflict evidence synthesis are analyzed, and the approach provided herein improving the accuracy and efficacy of integration is thereby illustrated.