Characterization and algorithm of decision system´s core based discernibility matrix

Calculating the core of a decision information system is the start of information reduction and a key step of decision rule making. In this paper, we analyze essential characters of core attributes of decision information system according to rough set theory. Then researching the relationship between discernibility matrix´ single attribute element and a core attribute. As algorithms, which were given out by Skowron and Zhang, has highly computing complicacy for calculating the core of decision system based on discernibility matrix. This paper gives out an improved algorithm, and proves it to be right. The simulation experiments shows that the new algorithm´s calculating work will be reduced according to the proportion of inconsistent objects has risen.