Kernel-Based Methods for Hypothesis Testing: A Unified View

Zaid Harchaoui (1) , Francis Bach (2, 3) , Olivier Cappé (4) , Eric Moulines (4)

1 LEAR - Learning and recognition in vision
2 LIENS - Laboratoire d'informatique de l'école normale supérieure
3 SIERRA - Statistical Machine Learning and Parsimony
4 LTCI - Laboratoire Traitement et Communication de l'Information

Kernel-based methods provide a rich and elegant framework for developing nonparametric detection procedures for signal processing. Several recently proposed procedures can be simply described using basic concepts of reproducing kernel Hilbert space embeddings of probability distributions, namely mean elements and covariance operators. We propose a unified view of these tools, and draw relationships with information divergences between distributions.

Domaines: Machine Learning [stat.ML]; Théorie [stat.TH]; Traitement du signal et de l'image [eess.SP]; Traitement du signal et de l'image [eess.SP]

Zaid Harchaoui, Francis Bach, Olivier Cappé, Eric Moulines. Kernel-Based Methods for Hypothesis Testing: A Unified View. IEEE Signal Processing Magazine, 2013, Special Issue on Advances in Kernel-Based Learning for Signal Processing, 30 (4), pp.87-97. ⟨10.1109/MSP.2013.2253631⟩. ⟨hal-00841978⟩