摘要
针对大众标注系统中标签语义模糊、标签检索效果不理想等难题,利用改进的潜在语义标引LSI提高标签检索效率。文章首先设计基于改进的LSI标签语义检索书目系统模型,定义新的标签-图书矩阵权重计算算法,并对关键技术进行详细说明,最后实现标签语义检索书目原型系统,并以"豆瓣读书"数据设计实验,取得了较好的实验效果。
Due to the tag semantic ambiguity and unsatisfactory tag retrieval from folksonomy system, this paper applies the improved latent semantic indexing to improve tag retrieval efficiency. Firstly, the authors design the tag semantic retrieval model of bibliography system based on the improved LSI, define a new weight algorithm of tag-book matrix, describe in detail key technologies, and finally implement a prototype system. Experimental results based on douban reading data show that the system achieves a good effect.
出处
《图书馆学研究》
CSSCI
北大核心
2014年第11期67-72,共6页
Research on Library Science
基金
国家社会科学基金项目"数字图书馆标签系统的语义挖掘研究"(项目编号:12CTQ003)的研究成果之一
