Information, Divergence and Risk for Binary Experiments Export

@article{citeulike:3851922, abstract = {We unify f-divergences, Bregman divergences, surrogate loss bounds (regret bounds), proper scoring rules, matching losses, cost curves, ROC-curves and information. We do this by systematically studying integral and variational representations of these objects and in so doing identify their primitives which all are related to cost-sensitive binary classification. As well as clarifying relationships between generative and discriminative views of learning, the new machinery leads to tight and more general surrogate loss bounds and generalised Pinsker inequalities relating f-divergences to variational divergence. The new viewpoint illuminates existing algorithms: it provides a new derivation of Support Vector Machines in terms of divergences and relates Maximum Mean Discrepancy to Fisher Linear Discriminants. It also suggests new techniques for estimating f-divergences.}, archivePrefix = {arXiv}, author = {Reid, Mark D. and Williamson, Robert C.}, citeulike-article-id = {3851922}, citeulike-linkout-0 = {http://arxiv.org/abs/0901.0356}, citeulike-linkout-1 = {http://arxiv.org/pdf/0901.0356}, citeulike-linkout-2 = {http://conflate.net/inductio/2009/01/information-divergence-and-risk/}, day = {5}, eprint = {0901.0356}, keywords = {classification, divergence, estimation, information, mine, probability, theory}, month = {Jan}, posted-at = {2009-01-06 07:33:17}, priority = {0}, title = {Information, Divergence and Risk for Binary Experiments}, url = {http://arxiv.org/abs/0901.0356}, year = {2009} }

TY - JOUR ID - citeulike:3851922 L3 - citeulike-article-id:3851922 N2 - We unify f-divergences, Bregman divergences, surrogate loss bounds (regret bounds), proper scoring rules, matching losses, cost curves, ROC-curves and information. We do this by systematically studying integral and variational representations of these objects and in so doing identify their primitives which all are related to cost-sensitive binary classification. As well as clarifying relationships between generative and discriminative views of learning, the new machinery leads to tight and more general surrogate loss bounds and generalised Pinsker inequalities relating f-divergences to variational divergence. The new viewpoint illuminates existing algorithms: it provides a new derivation of Support Vector Machines in terms of divergences and relates Maximum Mean Discrepancy to Fisher Linear Discriminants. It also suggests new techniques for estimating f-divergences. TI - Information, Divergence and Risk for Binary Experiments KW - classification KW - divergence KW - estimation KW - information KW - mine KW - probability KW - theory AU - Reid, Mark AU - Williamson, Robert PY - 2009/Jan/05/ UR - http://arxiv.org/abs/0901.0356 ER -