Using evolutionary multiobjective techniques for imbalanced classification data

Authors:
Sandra García;Ricardo Aler;Inés María Galván
Affiliations:
Computer Science Departament, Carlos III University of Madrid, Leganes, Spain;Computer Science Departament, Carlos III University of Madrid, Leganes, Spain;Computer Science Departament, Carlos III University of Madrid, Leganes, Spain
Venue:
ICANN'10 Proceedings of the 20th international conference on Artificial neural networks: Part I
Year:
2010

Citing 8
Cited 1

Neural networks: a systematic introduction

Neural networks: a systematic introduction
Neural Learning from Unbalanced Data

Applied Intelligence
Multi-class ROC analysis from a multi-objective optimisation perspective

Pattern Recognition Letters - Special issue: ROC analysis in pattern recognition
Classification by evolutionary ensembles

Pattern Recognition
The class imbalance problem: A systematic study

Intelligent Data Analysis
Multiobjective Problem Solving from Nature: From Concepts to Applications (Natural Computing Series)

Multiobjective Problem Solving from Nature: From Concepts to Applications (Natural Computing Series)
The use of the area under the ROC curve in the evaluation of machine learning algorithms

Pattern Recognition
A fast and elitist multiobjective genetic algorithm: NSGA-II

IEEE Transactions on Evolutionary Computation

A measure oriented training scheme for imbalanced classification problems

PAKDD'11 Proceedings of the 15th international conference on New Frontiers in Applied Data Mining

Quantified Score

Hi-index	0.00

Visualization

Abstract

The aim of this paper is to study the use of Evolutionary Multiobjective Techniques to improve the performance of Neural Networks (NN). In particular, we will focus on classification problems where classes are imbalanced. We propose an evolutionary multiobjective approach where the accuracy rate of all the classes is optimized at the same time. Thus, all classes will be treated equally independently of their presence in the training data set. The chromosome of the evolutionary algorithm encodes only the weights of the training patterns missclassified by the NN. Results show that the multiobjective approach is able to consider all classes at the same time, disregarding to some extent their abundance in the training set or other biases that restrain some of the classes of being learned properly.