Abstract: To realize the deep mining of diversification in information requirements among different user groups of enterprise information system,a log mining method by combining the theory of rough set and quotient space was put forward through comparison and analysis of the macrocosmic and microcosmic theoretical models of granular computing．Firstly,the users’ characteristic attributes table was established in terms of their use frequencies for different function modules of enterprise information system in a certain period and their working roles．Secondly,the decision tables in two user granular spaces were built by hierarchical division of user groups based on projection division method of quotient space theory．Furthermore,the corresponding user identification rules in two user granular spaces were respectively derived by means of applying knowledge acquisition method of rough set theory．Finally,the consistency and discrepancy of information requirements among different user groups was comprehensively analyzed from different angles．The application case shows the feasibility and effectiveness of the proposed method.

Key words: Rough set,Quotient space,Granular computing,Log mining,Enterprise information system,User granular space