SURVEY: A COMBINE APPROACH OF FEATURE SELECTION AND DIFFERENT CLUSTERING TECHNIQUE IN BREAST CANCER GENE DATA

DNA microarray datasets have large number of genes however only a small number of genes are required to detect a particular type of disease. So gene selection plays an important role in removing irrelevant features which improves accuracy. In this paper we discussed about feature extraction techniques like I - relief, PCA. As large microarray d atasets have the issue of dimensionality, PCA technique can be also used to reduce dimensions. After feature selection cluster analysis is performed to identify interesting patterns. Different types of clustering algorithms have been discussed like k - means , hierarchical, partitioning, model - based clustering and DB - Scan having their advantages and disadvantages in the result. Clustering validation techniques are discussed which can be used to calculate the exact number of clusters.