CSB '04 Proceedings of the 2004 IEEE Computational Systems Bioinformatics Conference
| Hi-index | 0.00 |
In this paper, we propose a new clustering procedure for high dimensional microarray data. Major difficulty in cluster analysis of microarray data is that the number of samples to be clustered is much smaller than the dimension of data which is equal to the number of genes used in an analysis. In such a case, the applicability of conventional model-based clustering is limited by the occurence of overlearning. A key idea of the proposed method is to seek a linear mapping of data onto the low-dimensional subspace before proceeding to cluster analysis. The linear mapping is constructed such that the transformed data successfully reveal clusters existed in the original data space. A clustering rule is applied to the transformed data rather than the original data. We also establish a link between this method and a probabilistic framework, that is, a penalized likelihood estimation of the mixed factors model. The effectiveness of the proposed method is demonstrated through the real application.