Research On Uncertainty And Knowledge Discovery Based On Evidence Theory

Rough set theory proposed by Pawlak, which can be used as an effective tool to deal with uncertainty and imprecise problems, all of the application is based on an equivalence relation. Dominance relation is a promotion of the equivalence relation, and the ordered information system is an extension of the classical information system. Compared with the equivalence relation, dominance relation can deal with some problems better in real life. In this paper, a novel method of evidence acquirement and combination based on rough set is proposed by introduced evidence theory into the ordered information system. Then discuss the uncertainty measures based on the different approximation operators. The main ideas in this article are listed in the following:1. From the point of evidence theory, combined with rough set theory, a new method and model of evidence acquirement and combination is proposed in ordered information system. Finally analyzed the rationality and effectiveness of the method proposed in this paper by an example analysis, and provided the advantages of this method compared with other models.2. Consider different definition of the upper and lower approximation operators, in this paper we investigated four types of lower and upper approximations and the corresponding uncertainty measures in ordered information system. Discussed the rationality of the corresponding uncertainty measures. Furthermore, the relationships among the four types lower and upper approximation operators and the corresponding uncertainty measures were got in the system. Finally, the four types of measures were tested to be correctness and rationality on some real-life data sets.