基于语义概念的海量短信文本聚类

提出一种基于语义概念的海量中文短信文本聚类方法。该方法从短信文本出发，利用《现代汉语语义分类词典》的级类主题词，在短信文本向量集中提取概念元组，形成表示聚类结果的高层概念，基于这些高层概念进行样本划分，从而完成整个聚类过程。实验结果表明，该聚类算法有较好的聚类结果且执行效率较高。

Clustering on Mass Chinese Short Message Text Based on Semantic Concept

This paper puts forward a clustering on mass Chinese short message texts technique based on semantic concept. The method starts from the Chinese short message itself, it uses classified hierarchy Subject Word of Thesaurus of Modern Chinese, then abstracts conceptional tuple from short message text, forms clustering result expressed by high-leveled concept, based on which these samples are divided, thus the whole clustering process is completed. Experimental result shows the clustering algorithm achieves a satisfactory clustering result and a better executive efficiency as well.