Clustering analysis of cancerous microarray data

Due to rapid advancement in microarray technology it is possible to measure expression of tens of thousands of genes simultaneously and as a result we have flood of data that need to be analyzed for the discovery of fruitful knowledge. Clustering is a well-known unsupervised learning approach that clubs a set of similar objects in groups that forms clusters. Cancerous microarray data may reveal fruitful information related to underlying mechanisms of cancer at molecular level which can be used for better diagnosis and therapies of cancers. In this paper, we applied four different clustering techniques, such as k-means, hierarchical, density-based and expectation maximization approaches, on five different kinds of cancerous gene expression data (lung, breast, colon, prostate, breast and ovarian cancer) for their analysis.