A heuristic method to attribute reduction for concept lattice
Junhong Wang, Jiye Liang, Yuhua Qian · 2010
Attribute reduction plays an important role in data mining based on concept lattice theory, which makes knowledge discovery from data easier and knowledge representation simpler. However, many existing methods often employ a discernibility matrix to calculate a set of complete reducts and are time-consuming. To solve this problem, in this paper, based on an expanded concept lattice, called closed label lattice, a heuristic algorithm to attribute reduction is proposed. The proposed algorithm and an illustrative example show that the algorithm can effectively obtain an attribute reduct from a formal context.