基于搜索引擎的文本主题识别  

Topic Detection Technology for Text Based on Search Engine

作　　者：石琼[1] 潘广贞[1] 

机构地区：[1]中北大学电子与计算机科学技术学院,山西太原030051

出　　处：《装备制造技术》2009年第6期92-94,共3页Equipment Manufacturing Technology

摘　　要：基于Google搜索引擎获取的结果,并在比较分析各种常用聚类算法的基础上,提出了一种基于Web文档的快速聚类算法,并采用多线程技术加以实现;该方法在保证聚类精度的前提下,提高了文本聚类的速度,适用于对大规模数据进行聚类;实验结果表明,该算法的聚类速度与文档的数量满足线性关系,优于各种常用聚类算法。Based on the Google＇s searching results, this article through compare and analysis all kinds of commonly used clustering algorithms, put forward a fast clustering algorithm for web documents, and it be realized by the technology of mnlfithreading. The method ensured the accuracy of cluster, improved the speed of the text clustering, and it applied to large-scale data＇s clustering. The results show that the method has reflected the linear relationship between clustering speed and the number of documents, and it is superior to all kinds of commonly used clustering algorithm.

关 键 词：搜索引擎 快速聚类算法 多线程 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]