A Classification Model using improved Hybrid Genetic Particle Swarm Optimization Algorithm based on Separability-Correlation Measure

Authors

  • B Renuka Devi
  • N. Sharmili
  • K. Vijaya Kumar
  • G. Jose Moses
  • Dr. E. Laxmi Lydia

Abstract

The unpredictable growth in information and data samples has engendered a crucial requirement for novel methodologies and mechanisms which can intellectually and spontaneously transform the processed information into valuable data and knowledge. Thus, it is very essential to carefully obtain the relevant information from the huge databases. Numerous techniques are already available in literature for mining of data. However, the Evolutionary Algorithm and Swarm Intelligent Approaches are playing a vital role in the form of extracting the relevant features from the database and supporting in constructing the classification Models. So as to further highlight the importance of both the approaches, in this paper, a methodology is presented that hybridized the Genetic Algorithm and Particle Swarm optimization for feature Selection by means of Separability-Correlation Measure. The experiment results shows that the proposed novel Feature Selection approach has a high global convergence possibility and a scarce average convergence iterations

 Keywords: Classification, Genetic Algorithm, Particle Swarm Optimization, Hybrid approach, Separability-Correlation Measure, Feature Selection

Downloads

Published

2020-01-06

Issue

Section

Articles