The research on application of granular transform in classification based on domain knowledge

Granular transform of original data, can obtain the userspsila expected results in their interested level, and can decrease the size of data at the same time. So, it has become an important research issue in data mining area. In this paper, a model of granular transform is constructed based on domain knowledge. Then, it is integrated with classification rules learning. A generalization algorithm is proposed to learning classification rules, which is focusing on granular level based on domain knowledge. By defining generation operators of classification and using concept lattice, this method chooses the most compact generalization level and path to generalize the attributes controlled by misclassification ratio based on certain mining tasks, and requests the userspsila interested knowledge from actual data in databases, in order to obtain the best classification rules. The experimentation shows the efficiency of this algorithm.