Title :
PAC-Bayesian learning with asymmetric cost (June 2011)
Author :
Llorens, Ashley J. ; Wang, I-Jeng
Author_Institution :
Appl. Phys. Lab., Johns Hopkins Univ., Laurel, MD, USA
Abstract :
PAC-Bayes generalization bounds offer a theoretical foundation for learning classifiers with low generalization error and predicting their performance on unseen data. Current formulations implicitly assume that the relative cost of misclassifying a positive or negative example is reflected by the class skew in the training dataset. We present a learning approach based on minimizing an asymmetric generalization bound that enables PAC-Bayesian model selection under a class-specific performance constraint.
Keywords :
Bayes methods; belief networks; learning (artificial intelligence); pattern classification; PAC-Bayesian model selection learning; asymmetric generalization minimization; class skew; data prediction; generalization error; learning classifier; probably approximately correct-Bayesian model selection learning; training dataset; Classification algorithms; Kernel; Machine learning; Minimization; Support vector machines; Training; Training data; PAC-Bayes; asymmetry; generalization bounds; kernel machines; neyman-pearson;
Conference_Titel :
Statistical Signal Processing Workshop (SSP), 2011 IEEE
Conference_Location :
Nice
Print_ISBN :
978-1-4577-0569-4
DOI :
10.1109/SSP.2011.5967816
