摘要
词义消歧在机器翻译、信息检索、语音语义识别等方面具有重要作用。为提高消歧质量,细化特征粒度,提出一种多特征词义消歧方案。通过依存句法分析提取上下文中多义词及义项的词性、依存结构、依存词等特征,细化特征粒度,并根据多特征构造权值函数,选择权值最大的义项作为多义词的义项。实验结果表明,与单一特征词义消歧相比,采用依存句法分析的多特征词义消歧方案细化了特征粒度,提高了消歧准确率。
Word Sense Disambiguation（ WSD） plays an important role in machine translation,information retrieval and speech semantic recognition. In order to improve the quality of disambiguation and refine the feature,a multi-feature granularity WSD scheme is proposed. The extraction of parts of speech,dependency structure and dependent words is used to detail feature grain by dependency parsing. The weight function is constructed according to the multiple features as the classifier, and the meaning with the largest weight is chosen as the sense of the polysemous word. Experimental results show that compared with single feature WSD,the multi-feature WSD scheme based on dependency parsing refines the feature and improves the accuracy of disambiguation.
出处
《计算机工程》
CAS
CSCD
北大核心
2017年第9期210-213,共4页
Computer Engineering
基金
国家自然科学基金(61272540)
关键词
词义消歧
依存句法
细化特征
多特征
权值
Word Sense Disambiguation（WSD）
dependency parsing
detailed feature
mulit-feature
weight
作者简介
史兆鹏（1990-）,男,硕士研究生,主研方向为自然语言处理、网络安全;E-mail：874257213@qq.com
硕士研究生
硕士研究生
