Efficient Multiclass Implementations of L1-Regularized Maximum Entropy

dc.creatorHaffner, Patrick
dc.creatorPhillips, Steven
dc.creatorSchapire, Rob
dc.date2005-06-29
dc.date.accessioned2026-07-07T03:23:11Z
dc.date.available2026-07-07T03:23:11Z
dc.descriptionThis paper discusses the application of L1-regularized maximum entropy modeling or SL1-Max [9] to multiclass categorization problems. A new modification to the SL1-Max fast sequential learning algorithm is proposed to handle conditional distributions. Furthermore, unlike most previous studies, the present research goes beyond a single type of conditional distribution. It describes and compares a variety of modeling assumptions about the class distribution (independent or exclusive) and various types of joint or conditional distributions. It results in a new methodology for combining binary regularized classifiers to achieve multiclass categorization. In this context, Maximum Entropy can be considered as a generic and efficient regularized classification tool that matches or outperforms the state-of-the art represented by AdaBoost and SVMs.
dc.description13 pages, describes new conditional maxent algorithm, to be submitted
dc.identifierhttps://arxiv.org/abs/cs/0506101
dc.identifierhttp://arxiv.org/abs/cs/0506101
dc.identifier.urihttp://salesiana.dossiersoluciones.com/handle/123456789/32845
dc.subjectMachine Learning
dc.subjectComputation and Language
dc.titleEfficient Multiclass Implementations of L1-Regularized Maximum Entropy
dc.typetext

Files

Collections