Model selection in logistic regression using p-values and greedy search

Authors:
Jan Mielniczuk;Pawe$#322/ Teisseyre
Affiliations:
Institute of Computer Science, Polish Academy of Sciences, Warsaw, Poland;Institute of Computer Science, Polish Academy of Sciences, Warsaw, Poland
Venue:
SIIS'11 Proceedings of the 2011 international conference on Security and Intelligent Information Systems
Year:
2011

Citing 3
Cited 0

Genetic algorithms for outlier detection and variable selection in linear regression models

Soft Computing - A Fusion of Foundations, Methodologies and Applications
Regression Modeling Strategies

Regression Modeling Strategies
Predictive models for the breeder genetic algorithm i. continuous parameter optimization

Evolutionary Computation

Quantified Score

Hi-index	0.00

Visualization

Abstract

We study new logistic model selection criteria based on p-values. The rules are proved to be consistent provided suitable assumptions on design matrix and scaling constants are satisfied and the search is performed over the family of all submodels. Moreover, we investigate practical performance of the introduced criteria in conjunction with greedy search methods such as initial ordering, forward and backward search and genetic algorithm which restrict the range of family of models over which an optimal value of the respective criterion is sought. Scaled minimal p-value criterion with initial ordering turns out to be a promising alternative to BIC.