基于数据分布特征的网页聚类算法

摘要

This paper proposed a new method of webpage clustering based on data distribute feature. The influence function be-tween data is computed by the affect of data to space, the most feature point is computed from bottom to top. The category is computed by the method from top to bottom. Experiments showed that the proposed webpage clustering based on data distribute feature is effective.%该文提出了一种基于数据分布特征的网页聚类算法。根据每一个数据点对整个集合空间的影响力，计算数据之间的影响函数，然后计算局部最大特征点，最后划分出类别。实验结果表明，该网页聚类算法是可行的。