Fuzzy Rough Incremental Attribute Reduction Applying Dependency Measures

Yangming Liu, Suyun Zhao, Hong Chen, Cuiping Li, Yanmin Lu

Abstract: Since data increases with time and space, many incremental rough based reduction techniques have been proposed. In these techniques, some focus on knowledge representation on the increasing data, some focus on inducing rules from the increasing data. Whereas there is less work on incremental feature selection (i.e., attribute reduction) from the increasing data, especially the increasing real valued data. And fuzzy rough sets is then applied in this incremental method because fuzzy rough set can effectively reduce attributes from the real valued data. By analyzing the basic concepts, such as lower approximation and positive region, of fuzzy rough sets on incremental datasets, the incremental mechanisms of these concepts are then proposed. An incremental algorithm is then designed. Finally, some numerical experiments demonstrate that the incremental algorithm is effective and efficient compared to non-incremental attribute reduction algorithms, especially on the datasets with large number of attributes.

External IDs:dblp:conf/apweb/LiuZCLL17