Unconfused ultraconservative multiclass algorithms

Ugo Louche 1 Liva Ralaivola 1
1 QARMA - éQuipe AppRentissage et MultimediA [Marseille]
LIF - Laboratoire d'informatique Fondamentale de Marseille
Abstract : We tackle the problem of learning linear classifiers from noisy datasets in a multiclass setting. The two-class version of this problem was studied a few years ago where the proposed approaches to combat the noise revolve around a Per-ceptron learning scheme fed with peculiar examples computed through a weighted average of points from the noisy training set. We propose to build upon these approaches and we introduce a new algorithm called UMA (for Unconfused Multiclass additive Algorithm) which may be seen as a generalization to the multiclass setting of the previous approaches. In order to characterize the noise we use the confusion matrix as a multiclass extension of the classification noise studied in the aforemen-tioned literature. Theoretically well-founded, UMA furthermore displays very good empirical noise robustness, as evidenced by numerical simulations conducted on both synthetic and real data.
Document type :
Journal articles
Complete list of metadatas

https://hal.archives-ouvertes.fr/hal-01146038
Contributor : Ugo Louche <>
Submitted on : Monday, April 27, 2015 - 3:23:41 PM
Last modification on : Monday, March 4, 2019 - 2:04:22 PM
Long-term archiving on: Wednesday, April 19, 2017 - 7:36:16 AM

Files

unconfused.pdf
Files produced by the author(s)

Licence


Distributed under a Creative Commons Attribution - ShareAlike 4.0 International License

Identifiers

Collections

Citation

Ugo Louche, Liva Ralaivola. Unconfused ultraconservative multiclass algorithms. Machine Learning, Springer Verlag, 2015, Machine learning, 99 (2), pp.351. ⟨10.1007/s10994-015-5490-3⟩. ⟨hal-01146038⟩

Share

Metrics

Record views

213

Files downloads

171