关键词: 搜索结果聚类, 词间语义相关度, 文档相似度, 主题词

Abstract:

Search engine results clustering has a high application value to the search service quality and the intelligence level, which clusters by seeking the document relations from the title and the document segment information. An algorithm based on topic words matching frequency is proposed. It automatically generate categories according to the graph of the semantic relevance, with other results clustering by topic words matching frequency. Experiments show that, compared with the STC and LINGO algorithms, the algorithm performs better.

Key words: search results clustering;semantic relevance between words;document similarity;topic word