CAJ | 학술논문

提出了词语相关度模型,作为在数据空间中发现数据源内容关联的一个基础。本模型基于HowNet,可以计算同种词性以及不同词性之间的相关度,融合了词语的相似度、关联度和实例因素,综合获得词语的内在相关性。通过对比实验发现,本模型所计算的词语相关度值更加符合人们主观上对词语相关性的认识。
제출료사어상관도모형,작위재수거공간중발현수거원내용관련적일개기출。본모형기우HowNet,가이계산동충사성이급불동사성지간적상관도,융합료사어적상사도、관련도화실례인소,종합획득사어적내재상관성。통과대비실험발현,본모형소계산적사어상관도치경가부합인문주관상대사어상관성적인식。
It is the basic of creating index,browsing,searching,querying and other services.The current search mostly is in the premise of having been obtained the relationship between the data.But it often is limited.The model of words relation proposed by this paper is the basic research of discovering the association of data source in the dataspace.The model of word relation is in view of HowNet,could compute the same and the different property of the words.It merges the word similarity,word association,the example factor and so on.Comprehensive these factors and receive the inner relation of words.Via the contract experiments,we find the model of words relation proposed by the paper is closed to and fit the cognition of words relation which people realize subjectively.