Abstract

thispaper presents a detailed study of different graph theoretic method i.e.clustering algorithm. A cluster is collection or group of data objects that aresimilar to each other with the same cluster object and not similar withother  Cluster object. Also it is studyon different feature selection algorithm. To overcome the limitations ofexisting algorithm. A feature selection may be related with both the efficiencyand effectiveness point of view.  FASTalgorithm is proposed. Features are different cluster relatively independent.Clustering based strategy has high probability of producing a subset ofimportant and independent features. To adopt the efficiency of fast clusteringfeature selection algorithm. It creates efficient minimum spanning treeclustering method.