Evolving data sets to highlight the performance differences between machine learning classifiers

Authors:
Thomas Raway;David J. Schaffer;Kenneth J. Kurtz;Hiroki Sayama
Affiliations:
Binghamton University, Binghamton, NY, USA;Binghamton University, Binghamton, NY, USA;Binghamton University, Binghamton, NY, USA;Binghamton University, Binghamton, NY, USA
Venue:
Proceedings of the 14th annual conference companion on Genetic and evolutionary computation
Year:
2012

Citing 4
Cited 0

Complexity Measures of Supervised Classification Problems

IEEE Transactions on Pattern Analysis and Machine Intelligence
On Classifier Domains of Competence

ICPR '04 Proceedings of the Pattern Recognition, 17th International Conference on (ICPR'04) Volume 1 - Volume 01
Microarray data mining with visual programming

Bioinformatics
No free lunch theorems for optimization

IEEE Transactions on Evolutionary Computation

Quantified Score

Hi-index	0.00

Visualization

Abstract

We present a preliminary study to evolve data sets that maximize performance differences between multiple machine learning classifiers. The aim is to provide useful information towards the decision of which machine learning classifier to use given a particular data set. While literature already exists on comparing multiple classifiers across multiple pre-existing data sets, our approach is novel and unique in that we evolved completely new data sets designed to highlight the performance differences between supervised learning classifiers. By investigating these evolved data sets, we hope to add to the knowledge base concerning which classifiers are appropriate for specific real world classification tasks.