基于错误驱动算法组合分类器及其在问题分类中的应用.pdf

ISSN 计算机研究与发展 1000—1239／CN11-1777／TP of Researchand JournalComputer Development 45(3)：535～541，2008 基于错误驱动算法组合分类器及其在问题分类中的应用 李 鑫 黄萱菁 吴立德 (复旦大学计算机科学系 上海200433) (1ixin@fudan．edu．on) CombinedClassifiersBasedonTBL andTheir in Multiple Algorithm Application Classification Question Li Wu Lide Xin，HuangXuanjing，and (Department Science，Fudan 200433) ofComputer University，Shanghai AbstractAsa activebranch ofnatural very languageprocessing，open-domainquestionanswering(QA) has beenattached attentionitcanunderstandthe innatural system increasing to，for question language， andthus itsuserswith andexact the provide compact results．Question intoseveralsemantic fora questions important and categories，isvery questionansweringsystemdirectly affectsthe ofthe in answers．Itsmain to QA correct taskis understandthe performancesystemselecting demandofusers．In this automatic classification paper，toinvestigate questionclassification，different features，suchas fromWordnetand structurefrom Bag—of—words，Bi—gram，synsetdependency Minipar， are vector suchmachine ensemble as compared．Supportmachine(SVM)and learning approaches transformation-basederror—driven andback artificialneuralnetwork learning(TBL