PSO Search-based Feature-selection Method for High Dimensional Data
M. Sathya1, S. Manju Priya2
1M. Sathya, Research Scholar, Department of Computer Science, Karpagam Academy of Higher Education, Coimbatore (Tamil Nadu), India.
2Dr. S. Manju Priya, Associate Professor, Department of CS, CA & IT, Karpagam Academy of Higher Education, Coimbatore (Tamil Nadu), India.
Manuscript received on 24 April 2019 | Revised Manuscript received on 02 May 2019 | Manuscript Published on 08 May 2019 | PP: 485-488 | Volume-7 Issue-5S3 February 2019 | Retrieval Number: E11860275S19/19©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: The growth of gene expression data from various techniques continues to expand. Addressing problems associated with high dimensional data and selecting relevant features have become more essential. Selection of relevant genes allows researchers to computationally explore gene expression to find functional genes, disease-causing genes and drug interactions to target specific genes. In the data mining community several feature-selection methods and techniques are continuously being studied and introduced. Selecting the best feature ranking method is still challenging. Feature-selection methods combine search methods and feature evaluation to find relevant features. The choice of search method has a significant relationship with feature ranking scores. In this paper, a new feature-selection method using PSO search strategy to derive high ranking feature subset is introduced. The extracted feature subset is experimentally studied on classification of Colon tumor using Colon dataset. The findings of the study show that PSO based search strategy shows better results than other methods. The study concludes that the proposed method can be used for high dimensional and classification problems on microarray dataset.
Keywords: Microarray Dataset, Feature-selection, Classification, Search Strategy, Feature Ranking, Particle Swarm Optimization, High Dimensional Problem.
Scope of the Article: Data Mining Methods, Techniques, and Tools