基于K均值和aiNet的两阶段文本聚类算法  

The Two-stage Text Clustering Algorithm Based on K-mesans and aiNet

作　　者：向永生[1] 刘燕婷[1] 徐家宁[1] 

机构地区：[1]长沙理工大学城南学院,长沙410076

出　　处：《微计算机信息》2009年第30期186-187,181,共3页Control & Automation

基　　金：2009年湖南省科技厅资助项目(2009GK3160)

摘　　要：为了克服基于aiNet的文本聚类算法在处理高维数时性能下降的缺点,设计一种基于聚类中心的虚拟坐标映射机制的文本表示模型,实现文本向量模型的降维技术。在文本进行K均值聚类后,对文本向量进行降维,再将其作为aiNet文本聚类的输入。实验表明,通过两阶段的算法对文本进行聚类,有效提高了aiNet文本聚类质量。In order to overcome the shortcomings of performance-drop of text clustering algorithm based-on aiNet in dealing with high-dimensional data, designed an text expressive model which is based on cluster centers with virtual coordinate mapping mechanism, to drop text vector dimensions. High-dimensional text data using text expressive model after clustering by K-means,which were taken as input data for clustering algorithm based-on aiNet. The experimental results show that the quality of cluster results have improved effectively by two-stage cluster text algorithm.

关 键 词：文本聚类 向量空间模型 人工免疫网 K均值聚类算法 

分 类 号：TP31[自动化与计算机技术—计算机软件与理论]