Article ID Journal Published Year Pages File Type
402827 Knowledge-Based Systems 2013 8 Pages PDF
Abstract

Decision rule mining is an important technique in many applications. In this paper, we propose a new rough set approach for rule induction based on a significance measure, called classification consistency rate. The approach implements the rule induction from the viewpoint of attribute rather than descriptor. The proposed algorithm is tested and compared with LEM2 algorithm on several real-life data sets added with different levels of inconsistent data. The results show that the proposed algorithm is effective in rule induction for inconsistent data.

Related Topics
Physical Sciences and Engineering Computer Science Artificial Intelligence
Authors
, , , ,