Concept Lattice-based Classification Rule Extraction Algorithm

Concept lattice is an effective tool for knowledge discovering and data mining,and classification rule mining is an important task in data mining. For classification task, a batch constructing algorithm of concept lattice oriented classification and a classification rule mining algorithm based on the concept lattice and conditional entropy are presented, so that the constructing efficiency of concept lattice and classification efficiency of the classification rules are improved. The main research works are as follows:(1) A batch constructing algorithm of concept lattice oriented classification is presented. For classification rule acquisition, the algorithm only generates the concept lattice nodes in which the classification attributes are contained during constructing concept lattice. So that the compare operation of concept intent can be decreased, increase the constructing efficiency of concept lattice is improved. Finally, the experiment results show the correctness and the validity of the algorithm by taking UCI data as the formal context.(2) A classification rule acquisition algorithm based on the concept lattice and conditional entropy is presented. First, extent support coefficient is used to depicting the important of the intent. Second, the rule order in the classification rule set is arranged by using conditional entropy. In the end, the experimental results show the algorithm can effectively improve the classification efficiency under the classification accuracy unchanged by taking UCI data as the formal context.