[1]郑新立 徐云青 骆昌日[].LSI模型在信息检索中的应用[J].计算机技术与发展,2006,(10):160-162.
ZHENG Xin-li,XU Yun-qing,LUO Chang-ri,et al.Application of LSI Model in Chinese Information Retrieval[J].,2006,(10):160-162.
点击复制
LSI模型在信息检索中的应用(
)
《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]
- 卷:
-
- 期数:
-
2006年10期
- 页码:
-
160-162
- 栏目:
-
应用开发研究
- 出版日期:
-
1900-01-01
文章信息/Info
- Title:
-
Application of LSI Model in Chinese Information Retrieval
- 文章编号:
-
1673-629X（2006）10-0160-03
- 作者:
-
郑新立1 徐云青2 骆昌日[3]
-
[1]国防科技大学计算机学院[2]衢州学院[3]华中师范大学网络教育学院
- Author(s):
-
ZHENG Xin-li; XU Yun-qing; LUO Chang-ri
-
[1]School of Computer, National University of Defence Technology[2]Quzhou College[3]School of Network Education, Central China Normal University
-
- 关键词:
-
LSI; 中文信息检索; SVD
- Keywords:
-
LSI; Chinese information retrieval; SVD
- 分类号:
-
TP391.3
- 文献标志码:
-
A
- 摘要:
-
介绍了LSI技术在中文信息中的应用及其在实际应用中可能会出现的问题，并给出了解决办法。通过将查询和文档在多维空间进行投影，将共现词投到同一维上，建立“词-文档”高维空间，使用SVD方法对高维空间进行分解、降维，得到词、文档、查询的向量表示，达到对中文信息的检索；使用PRF（Pseudo Relevance Feedback）技术对其进行改进
- Abstract:
-
Presents the application and the problems in application about the LSI techniques. Then give the solution about the problems. LSI technique projects queries and documents make into a high dimensions space with latent semantic, Co-occurring terms are projected onto the same dimensions, non - co- occurring terms are projected onto different dimensions, SVD method was employed to decompose the original “term - document”matrix reduce its dimensions and get the representation of queries, terms and documents in the reduced dimension space for Chinese information retrieval. And use PRF technique to improve the performance of retrieval
备注/Memo
- 备注/Memo:
-
邵新立（1968～），男，浙江人，硕士研究生，研究方向为信息检索
更新日期/Last Update:
1900-01-01