Abstract—Rough set theory is an effective mathematical tool to process inaccurate, inconsistent and incomplete information. The primary goal of rough set theory has been outlined as a classificatory analysis of data: given a data table, rough set algorithms induce a set of relevant concepts such as rules providing a classification of data. However, these concepts may contradict with some priori knowledge in expert system which causes problem in reasoning. This paper proposes a classification system based on rough set theory and non-revision reasoning which tolerates the inconsistency between the generated concepts and priori knowledge. This approach integrates knowledge from multi-sources without data normalizing, which improves the efficiency and the rationality of the classification result. Moreover, integration of knowledge instead of data also preserves the information privacy and reduces the cost on transfer.
Index Terms—Classification, decision rule, non-revision, rough set.
The authors are with Dalian Maritime University Dalian 116026, PRC (e-mail: email@example.com, firstname.lastname@example.org, email@example.com).
Cite: Zhang Song, Deng Ansheng, and Qu Yanpeng, "Classification System Based on Non-Revision Reasoning and Rough Set Theory," International Journal of Machine Learning and Computing vol.6, no. 1, pp. 1-8, 2016.