社会标注系统自适应网页聚类算法研究

Research on Adaptive Webpage Clustering Algorithm in Social Annotation System

文中提出了一种社会标注系统自适应网页聚类算法,可以自适应找出类别数目k并完成聚类。将随机选择的15组网页语料进行聚类对比实验,采用准确率、召回率、F-Measure值、Purity和NMI这5个指标来评测聚类的效果。实验结果表明,文中所提出的自适应网页聚类算法效果较好。

This paper presents adaptive webpage clustering algorithm in social annotation system,it can find class number k adaptively and complete clustering. It randomly selects 15 groups of Webpage corpus for contrast clustering experiment,and uses the accuracy rate,recall rate,F-Measure value,Purity and NMI indicators to evaluate the clustering effect. Experimental results show that the proposed adaptive webpage clustering algorithm in social annotation system is effective.