基于BBS语料的话题提取主要是从大量的BBS论坛讨论信息中,将正在或近期讨论的各种话题提取出来。在自主开发的一套话题提取系统中采用了一个原始聚类算法,能够对真实的BBS语料进行有效话题提取。随后将语料中的关联信息引入到原始聚类算法中进行改进,提高了算法的性能,取得了良好的效果。
Topic detection and tracking in BBS is mainly to detect the topics being discussed or had been discussed recently from a host of related information from BBS. The topic detection system with a clustering algorithm is effective for topic detection and tracking in BBS. Im- provement is made on the related information of BBS,which is introduced to the original clustering algorithm. The performance of the algorithm is improved, and better results are achieved.