Journal
INTERNATIONAL JOURNAL OF MACHINE LEARNING AND CYBERNETICS
Volume 1, Issue 1-4, Pages 63-74Publisher
SPRINGER HEIDELBERG
DOI: 10.1007/s13042-010-0008-6
Keywords
Cancer recognition; Gene selection; Neighborhood mutual information; Maximal relevancy; Minimal redundancy
Categories
Funding
- National Natural Science Foundation of China [60703013, 61070089]
Ask authors/readers for more resources
Gene selection is a key problem in gene expression based cancer recognition and related tasks. A measure, called neighborhood mutual information (NMI), is introduced to evaluate the relevance between genes and related decision in this work. Then the measure is combined with the search strategy of minimal redundancy and maximal relevancy (mRMR) for constructing a NMI based mRMR gene selection algorithm (NMI_mRMR). In addition, it is also found that the first k best genes with respect to NMI are usually enough for cancer classification. We can just perform mRMR on these genes and remove the rest in the preprocessing step, which will lead to reduction of computational time. Based on this observation, an efficient gene selection algorithm, denoted by NMI_EmRMR, is proposed. Several cancer recognition tasks are gathered for testing the proposed technique. The experimental results show NMI_EmRMR is effective and efficient.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available