Using feature selection approaches to find the dependent features

Authors:
Qin Yang;Elham Salehi;Robin Gras
Affiliations:
School of Computer Science, University of Windsor, Windsor, ON, Canada;School of Computer Science, University of Windsor, Windsor, ON, Canada;School of Computer Science, University of Windsor, Windsor, ON, Canada
Venue:
ICAISC'10 Proceedings of the 10th international conference on Artificial intelligence and soft computing: Part I
Year:
2010

Citing 9
Cited 2

Neural network design

Neural network design
Wrappers for feature subset selection

Artificial Intelligence - Special issue on relevance
Feature subset selection by Bayesian network-based optimization

Artificial Intelligence
Induction of Decision Trees

Machine Learning
Naive (Bayes) at Forty: The Independence Assumption in Information Retrieval

ECML '98 Proceedings of the 10th European Conference on Machine Learning
Correlation-based Feature Selection for Discrete and Numeric Class Machine Learning

ICML '00 Proceedings of the Seventeenth International Conference on Machine Learning
Consistency-based search in feature selection

Artificial Intelligence
Improvements to Platt's SMO Algorithm for SVM Classifier Design

Neural Computation
The WEKA data mining software: an update

ACM SIGKDD Explorations Newsletter

Correlation between genetic diversity and fitness in a predator-prey ecosystem simulation

AI'11 Proceedings of the 24th international conference on Advances in Artificial Intelligence
The emergence of new genes in ecosim and its effect on fitness

SEAL'12 Proceedings of the 9th international conference on Simulated Evolution and Learning

Quantified Score

Hi-index	0.00

Visualization

Abstract

Dependencies among the features can decrease the performance and efficiency in many algorithms. Traditional methods can only find the linear dependencies or the dependencies among few features. In our research, we try to use feature selection approaches for finding dependencies. We use and compare Relief, CFS, NB-GA and NB-BOA as feature selection approaches to find the dependent features among our artificial data. Unexpectedly, Relief has the best performance in our experiments, even better than NB-BOA, which is a population-based evolutionary algorithm that used the population distribution information to find the dependent features. It may be because some weak "link strengths" between features or due to the fact that Naïve Bayes classifier which is used in these wrapper approaches cannot represent the dependencies between features. However, the exact reason for these results still is an open problem for our future work.