A new approach to feature subset selection
Dazhong Liu, Zhi-Jing Feng, Xizhao Wang · 2005
The paper presents a brief overview to the approaches of feature subset selection (FSS), commonly used in machine learning or pattern recognition. A combined algorithm based on the two algorithms, i.e., the mutual information selector (MIFS) and relevance information selector (RELFSS), is put forward. Experiments show some advantages of the combined algorithm.