An Knowledge Reduction Algorithms in Data Mining Based on Rough Set Theory

Rough set theory, as a new mathematical tool for dealing with incompleteness and uncertainty of information, play an important role in recent data mining research. Knowledge reduction or attribute reduction is one of the core contents in rough set theory. It is also an important step in data mining. Due to the randomness of data collection in real life, there are many inconsistent decision tables in database. So some existing algorithms are no longer applicable and establishing a new algorithm suitable for both consistent and inconsistent decision tables is necessary. In the paper, an improved attribute reduction algorithm based on equivalence partition is proposed. Its main idea is to consider the decision table as a whole and form a new decision table based on equivalence partition. Then the reduction is put on this new decision table. In the paper it is proved that the attribute reduction and the core of this new decision table are the same as that of the original decision table. The algorithm is proved to be correct, effective and more applicable through theoretic proof and case analysis.