A Survey on Clustering Based Attribute Selection Algorithm for High Dimensional Data

Sonam R Yadav, Ravi P Patki

Abstract: Attribute selection includes recognizing a subset of the most useful attributes that delivers good results as the Original. Whole attribute selection algorithm highlight choice calculation may be assessed from both the efficiency and effectiveness perspectives. While the efficiency concerns the time needed to discover a subset of attributes, the effectiveness is identified with the quality of the subset of attributes. In this paper we discussed the survey on a clustering-based attribute selection algorithm. We also discussed about the FAST algorithm lives up to expectations in two stages. In the first step, attributes are separated into clusters by using graph-theoretic clustering methods. In the second step, the most illustrative attributes that is firmly identified with target classes is chosen from every clusters to structure a subset of attributes.

Keywords: Attribute subset selection, filter method, attribute clustering, and graph-based clustering