融入概念的高效主题分析框架与关键技术研究
Similar topics have higher probability values over the relevant words,Thus, a topic is near its similar topics in a probability vector space,Two different probability distributions might have the same label.Unsupervised topic modeling,Supervised topic modeling,Data-independent Hashing,Data-dependent Hashing.Unsupervised hashing,Supervised hashing,Distribution similarity ranking method.Traditional topic labeling, Select manually,Automatical labeling,hard for new emerging topics,time-consuming.
自然语言处理 主题分析 信息抽取
毛先领
北京理工大学
国内会议
第十五届中国自然语言处理青年学者研讨会(YSSNLP2018)
南京
中文
1-19
2018-05-01(万方平台首次上网日期,不代表论文的发表时间)