Abstract: Text clustering is subject to more and more attention as a hot research for a long time. This paper, according to the idea of ant colony algorithm, designs a text clustering algorithm based on ant colony algorithm, analyzes the structure of cluster and generation process, discusses the rules and algorithms of merging cluster in clustering. Experimental results prove that the method has good clustering effect, and effectively improve text’s recall rate of query.

Key words: text clustering, ant colony algorithm, cluster, similarity

摘要： 针对目前文本检索后的相关反馈信息较少用于文本聚类中的问题，根据蚂蚁觅食聚类算法的思想，将文本检索后的相关反馈信息应用到文本聚类过程中，提出一种基于蚁群算法的文本聚类算法。分析簇的结构及其生成过程，论述聚类中簇合并的规则及算法。实验结果表明，该算法具有良好的聚类效果，能有效提高查询的文本召回率。

关键词: 文本聚类, 蚁群算法, 簇, 相似度

CLC Number: 

• TP393