Research On Algorithm For Attribute Reduct & Core Computation Based On Rough Set Theory

Rough Set theory is founded by a Polish mathematician-Pawlak Z. in 1982, it is a new kinds of mathematical tool to deal with Vagueness and Uncertainty problems. The main advantage of Rough Set theory is that it has no use for any preliminary or additional information about data. The effective reduct algorithm is the foundation to use the rough set theory in data maining and knowledgediscovery in database. Moreover, the computation core is the key step in many heuristic reduct algorithm. so search a fast reduct algorithm and core computation algorithm are however main research issues of Rough Set theory.Some main problems are mainly researched in the paper as follows:(1) A heuristic attribute reduction algorithm based on conditional information entropy is proposed by utilizing the idea that the universe is changed gradually. The example shows that this algorithm is simple and efficient while the number of element is larger and the number of attribute is smaller in universe. These results will be helpful for knowledge acquiring based on rough set.(2) The changing mechanism of conditional entropy in decision table is analyzed, when a new object is added to the original universe. Based on the mechanism, a new incremental algorithm of the computation of core is provided. In view of information view, computation of new conditional entropy need only to find the condition class and the decision class with same value to the element added in this algorithm. This algorithm can discover incremental attribute core in decision table, and its validity have been depicted by using a concrete instance as well.Finally, achieved main results in the paper are generalized, theoretical significance and potential applied value of the research are explained, some problems farther needed and remarkable research direction for the future are indicated.