People’s Police Academy,

Attributes reduction methods based approach traditional rough sets perform on the decision tables with discrete attribute value domain. In fact the data is usually the real values or symbols should be reduced attributes of traditional rough sets proved ineffective because of its failure to preserve the difference of data on original objects. This problem is solved with the attributes reduction methods based approach fuzzy rough set to overcome the limitations of the method according to previous rough set approach. This paper improves, analyzes and evaluates two methods of attribute reduction based on the degree of dependence between attributes and discernibility matrix of fuzzy rough se