Publication Details

Constrained Classification of Large Imbalanced Data by Logistic Regression and Genetic Algorithm

HLOSTA Martin, STRÍŽ Rostislav, KUPČÍK Jan, ZENDULKA Jaroslav and HRUŠKA Tomáš. Constrained Classification of Large Imbalanced Data by Logistic Regression and Genetic Algorithm. International Journal of Machine Learning and Computing, vol. 2013, no. 3, pp. 214-218. ISSN 2010-3700. Available from: http://www.ijmlc.org/index.php?m=content&c=index&a=show&catid=36&id=304
Czech title
Klasifikace rozsáhlých nevyvážených dat pomocí logistické regrese a genetického algoritmu s omezujícími podmínkami
Type
journal article
Language
english
Authors
URL
Keywords

Imbalanced data, classification, genetic algorithm, logistic regression

Abstract

Imbalance in data classification is a frequently discussed problem that is not well handled by classical classification techniques. The problem we tackled was to learn binary classification model from large data with accuracy constraint for the minority class. We propose a new meta-learning method that creates initial models using cost-sensitive learning by logistic regression and uses these models as initial chromosomes for genetic algorithm. The method has been successfully tested on a large real-world data set from our internet security research. Experiments prove that our method always leads to better results than usage of logistic regression or genetic algorithm alone. Moreover, this method produces easily understandable classification model.

Published
2013
Pages
214-218
Journal
International Journal of Machine Learning and Computing, vol. 2013, no. 3, ISSN 2010-3700
Publisher
of Computer Science and Information Technology Press
BibTeX
@ARTICLE{FITPUB10277,
   author = "Martin Hlosta and Rostislav Str\'{i}\v{z} and Jan Kup\v{c}\'{i}k and Jaroslav Zendulka and Tom\'{a}\v{s} Hru\v{s}ka",
   title = "Constrained Classification of Large Imbalanced Data by Logistic Regression and Genetic Algorithm",
   pages = "214--218",
   journal = "International Journal of Machine Learning and Computing",
   volume = 2013,
   number = 3,
   year = 2013,
   ISSN = "2010-3700",
   language = "english",
   url = "https://www.fit.vut.cz/research/publication/10277"
}
Back to top