基于知网的词汇语义相似度计算方法研究

所需积分/C币:15 2018-04-28 11:29:50 423KB PDF
26
收藏 收藏
举报

利用知网的义原层次树,考虑树的深度、密度等因素对义原节点权重的影响,得到义原相似度
3331· 3 CPU 0.8974 0.6760 0.6760 DEF tool I wash I #clothing 1213 F tool alterForm level I 2 1 2 0.4444 0.6154 0.444 0.5000 2 实体 物质精神 property特性 newness旧…sex性别 goodbad好不 人匠稼树花草蔬菜水果 wdnm1男mddy园例性面窗画鱼 图1义原层次不同的树 图2义原密度不均的树 CPU DEF= part I DEF =part I lo animalHuman I heart I ef=part I %o fac flesh I # police I 0。 2 ?1994-2015 China Academic Journal Electronic Publishing House. All rig 4 rights sser y weight tever/ wwcnki.net 3332 27 sIm yi weight k simy;q=δ weight k x dep depth× depth+1 depth ∑imx;y Lin sIm (0= sIn xy∈L 6 east common node lcn log lcn log f 12× log lCn ht level i 2 log a log f 3.2 5 slm,红 mB yj=12…n m 3 1≤i≤m1≤j≤ X 8 Y y 图3集合X和P中元素构成的二部图 simConcept a h=∑:∏ b β1+β2+B3+B4=1 a y L ,≥B2≥B4 b s ?1994-2015cHinaAcademicJournalElectronicpUblishingHouse.Allrightsreservedhttp://www.cnki.net 3333· 358333、0.86250 “CPU”“” 0.5 6。6162 3B4 0.6、0.7、0.8、0.9 β1=0.5β2=0.2B3=0.1564=0.15 y1=0.8y2=0.2 δ=0.05 1 LEE L. Similarity -based approaches to natural language processing D. Cambridge Harvard University 1997 2 BROWN P. Word sense disambiguation using tactical methods C/i' 4 Proc of the 29th Meeting of Association for Computational I inguistics 0.722222 0.800000 0.80000 0.836250 0.6111 0.711250 0.2165 0.531018 0.800000 0.836250 5 RADA R MILI H BICKNELL E. Development and application of a 0.1l1628 0.097674 metric on semantic nets J. EEE Trans on System Man and CyberneTics 1989 19 1 17-30 6 LEE JH KIM MH LEE YI. Information retrieval based on con- ceptual distance in ISA hierarchies J. Journal of Documenta- 993492188-20 7 RESNIK P. Using information content to evaluate semantic similarity in a taxonomy C //Proc of International Joint Conference for Artifi cial intelli 1995448-453 8 AGIRRE E RIGAU G. A proposal for word sense disambiguation using conceptual distance C //Proe of International Conference on Recent Advances in Natural Language Processing. 1995 5 20082258489 CPU 0.897368 0.858333 0.676000 0.506240 0.862500 2009233116-120 0.722222 0.722222 0.850000 How Net eb/OL.2007-01-012009-12 P 0.052137 0.006240 0.873349 0.850000 ℃PU “CPU”、 200733 0.897368、 14 LIN De-kang. An information -theoretic definition of similarity seman- 0.722222 tic distance in WordNet C //Proc of the 15 th International Confer- nd3oz-t661G hina Academic Journal Electronic Publishing Hor Machine learning y San Francisco/ morgan ckauimaen Pub- 0

...展开详情
立即下载
限时抽奖 低至0.43元/次
身份认证后 购VIP低至7折
一个资源只可评论一次,评论内容不能少于5个字
您会向同学/朋友/同事推荐我们的CSDN下载吗?
谢谢参与!您的真实评价是我们改进的动力~
上传资源赚钱or赚积分
最新推荐
基于知网的词汇语义相似度计算方法研究 15积分/C币 立即下载
1/0