Evaluating Learning Algorithms to Construct Rule Evaluation Models Based on Objective Rule Evaluation Indices
Hidenao Abe, Shusaku Tsumoto, Miho Ohsaki, Takahira Yamaguchi · 2007
In the present paper, we describe an evaluation of our rule evaluation support method with constructive meta- learning scheme for post-processing of mined results with rule evaluation models based on objective indices. Postprocessing of mined results is one of the key processes in the data mining process. However, it is difficult for human experts to completely evaluate several thousand of rules from a large dataset with noises. To reduce the costs in such a rule evaluation task, we have developed a rule evaluation support method with rule evaluation models, which learn from objective indices for mined classification rules and evaluations by a human expert for each rule. To enhance the adaptability of rule evaluation models, we introduced a constructive meta-learning scheme to choose proper learning algorithms. Then, we performed the case study on the meningitis data mining as an actual problem. In addition, we evaluated the proposed method using the ten rule sets obtained from the ten UCI datasets. The obtained results demonstrate the applicability of the proposed rule evaluation support method.