Feature-wise Competitive Repetition Suppression Learning for Gene Data Clustering and Feature Ranking

The paper extends Competitive Repetition-suppression (CoRe) learning to deal with high dimensional data clustering. We show how CoRe can be applied to the automatic detection of the unknown cluster number and the simultaneous ranking of the features according to learned relevance factors. The effectiveness of the approach is tested on two freely available data sets from gene expression data and the results show that CoRe clustering is able to discover the true data partitioning in a completely unsupervised way, while it develops a feature ranking that is consistent with the state-of-the-art lists of gene relevance.