A Gaussian prior for smoothing maximum entropy models Export

@misc{citeulike:1613369, abstract = {In certain contexts, maximum entropy (ME) modeling can be viewed as maximum likelihood training for exponential models, and like other maximum likelihood methods is prone to overfitting of training data. Several smoothing methods for maximum entropy models have been proposed to address this problem, but previous results do not make it clear how these smoothing methods compare with smoothing methods for other types of related models. In this work, we survey previous work in maximum entropy...}, author = {Chen, S. and Rosenfeld, R.}, citeulike-article-id = {1613369}, citeulike-linkout-0 = {http://citeseer.ist.psu.edu/chen99gaussian.html}, citeulike-linkout-1 = {http://citeseer.lcs.mit.edu/chen99gaussian.html}, citeulike-linkout-2 = {http://citeseer.ifi.unizh.ch/chen99gaussian.html}, citeulike-linkout-3 = {http://citeseer.comp.nus.edu.sg/chen99gaussian.html}, keywords = {discriminative, lm, log-linear, machine-learning, maxent, non-local}, posted-at = {2007-10-01 20:38:16}, priority = {2}, title = {A Gaussian prior for smoothing maximum entropy models}, url = {http://citeseer.ist.psu.edu/chen99gaussian.html}, year = {1999} }

TY - GEN ID - citeulike:1613369 L3 - citeulike-article-id:1613369 N2 - In certain contexts, maximum entropy (ME) modeling can be viewed as maximum likelihood training for exponential models, and like other maximum likelihood methods is prone to overfitting of training data. Several smoothing methods for maximum entropy models have been proposed to address this problem, but previous results do not make it clear how these smoothing methods compare with smoothing methods for other types of related models. In this work, we survey previous work in maximum entropy... TI - A Gaussian prior for smoothing maximum entropy models KW - discriminative KW - lm KW - log-linear KW - machine-learning KW - maxent KW - non-local AU - Chen, S AU - Rosenfeld, R PY - 1999/// UR - http://citeseer.ist.psu.edu/chen99gaussian.html ER -