A study of network-based approach for cancer classification

The advent of high-throughput techniques such as microarray data enabled researchers to elucidate process in a cell that fruitfully useful for pathological and medical. For such opportunities, microarray gene expression data have been explored and applied for various types of studies e.g. gene association, gene classification and construction of gene network. Unfortunately, since gene expression data naturally have a few of samples and thousands of genes, this leads to a biological and technical problems. Thus, the availability of artificial intelligence techniques couples with statistical methods can give promising results for addressing the problems. These approaches derive two well known methods: supervised and unsupervised. Whenever possible, these two superior methods can work well in classification and clustering in term of class discovery and class prediction. Significantly, in this paper we will review the benefit of network-based in term of interaction data for classification in identification of class cancer.