Handling uncertainties in SVM classification

Emilie Niaf (1, 2) , Rémi Flamary (3) , Carole Lartizien (1) , Stéphane Canu (3)

1 CREATIS - Centre de Recherche en Acquisition et Traitement de l'Image pour la Santé
2 Applications des ultrasons à la thérapie
3 LITIS - Laboratoire d'Informatique, de Traitement de l'Information et des Systèmes

This paper addresses the pattern classification problem arising when available target data include some uncertainty information. Target data considered here is either qualitative (a class label) or quantitative (an estimation of the posterior probability). Our main contribution is a SVM inspired formulation of this problem allowing to take into account class label through a hinge loss as well as probability estimates using epsilon-insensitive cost function together with a minimum norm (maximum margin) objective. This formulation shows a dual form leading to a quadratic problem and allows the use of a representer theorem and associated kernel. The solution provided can be used for both decision and posterior probability estimation. Based on empirical evidence our method outperforms regular SVM in terms of probability predictions and classification performances.

Mots clés: support vector machines; maximal margin algorithm; uncertain labels.; uncertain labels

Domaines: Apprentissage [cs.LG]

Emilie Niaf, Rémi Flamary, Carole Lartizien, Stéphane Canu. Handling uncertainties in SVM classification. IEEE Workshop on Statistical Signal Processing, Jun 2011, Nice, France. ⟨hal-00582789v1⟩