Abstract | ||
---|---|---|
In this paper we compare classifier selection using cross-validation with meta-learning, using as meta-features both the cross-validation errors and other measures characterizing the data. Through simulation experiments we demonstrate situations where meta-learning offers better classifier selections than ordinary cross-validation. The results provide some evidence to support meta-learning not just as a more time efficient classifier selection technique than cross-validation, but potentially as more accurate. It also provides support for the usefulness of data complexity estimates as meta-features for classifier selection. |
Year | Venue | Keywords |
---|---|---|
2012 | Pattern Recognition | learning (artificial intelligence),pattern classification,cross-validation errors,cross-validation-based classifier selection improvement,data characterization,data complexity estimation,meta-learning features |
Field | DocType | ISSN |
Margin (machine learning),Pattern recognition,Computer science,Artificial intelligence,Classifier (linguistics),Cross-validation,Bayes classifier,Machine learning,Data complexity,Quadratic classifier | Conference | 1051-4651 |
ISBN | Citations | PageRank |
978-1-4673-2216-4 | 1 | 0.35 |
References | Authors | |
4 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Jesse H. Krijthe | 1 | 1 | 0.35 |
Tin Kam Ho | 2 | 1 | 0.35 |
Marco Loog | 3 | 1796 | 154.31 |