A Semantic Kernel for Computing Relatedness using Wikipedia
measuring the semantic relatedness between terms is a fundamental task of many natural language processing. In this paper, we propose a new kernel method with the aid of Wikipedia for this purpose, which first maps the Wikipedia concept into the Wikipedia category feature space, and then incorporates the semantic knowledge between categories into the kernel method. Specially, we implement the mapping step by just analyzing the internal hyperlinks that appear in the concepts article content and the corresponding categories of the concepts which the hyperlinks link to. The new approach is evaluated on a widely used dataset WordSimilarity-353, and the correlation value of our approach with the scores of human judgments is 0.5507, which outperforms many previous approaches.
semantic relatedness kernel method Wikipedia
Hongchun Zhang Tingting He
Department of Computer Science, HuaZhong Normal University,Network Media Branch, National Language R Department of Computer Science, HuaZhong Normal University, Network Media Branch, National Language
国际会议
重庆
英文
227-230
2011-01-21（万方平台首次上网日期，不代表论文的发表时间）