一种基于名词短语的检索结果多层聚类方法

为了对检索结果获取高质量的聚类效果,提取名词短语作为候选类别标签,根据候选类别标签分布情况生成基础类,再使用具有线性时间复杂度的一趟聚类算法对基础类进行多层聚类。与NEC,STC和Lingo算法的对比实验表明:该方法在类别标签的可读性、有效性以及聚类性能上都优于以上3种方法。

A multi-level clustering approach based on noun phrases for search results

In order to  get high qualitative clustering results, the noun phrases was selected as candidate cluster labels and generates basic clusters based on the distribution of candidate cluster labels. And then multi-level clustering was proceeded on basic clusters by using one pass clustering algorithm with linear time complexity. The comparative experiment was carried with our method, NEC algorithm, STC algorithm and Lingo  algorithm, and the results showed that our method could get more informative, readable cluster labels and more effective than other three methods.