摘要
对一种基于动态可调自组织神经网络(the dynamic adaptive self-organizing map neural network,简称DASOM)的增量中文文本聚类方法进行研究,认为其只需处理更新数据,提高聚类速度,并能自动抽取SOM聚类结果;DASOM模型具有动态的结构,通过数值实验表明该方法对中文文本增量聚类具有有效性。
This paper studies an incremental clustering method of Chinese documents based on dynamic adaptive self-organizing map neural network (in short DASOM),which can speed up clustering because it only recalculate the update datas. Features of DASOM are dynamic structure and organizing SOM's clusters results automatically. Numerical experiment shows that the method is efficient for clustering Chinese documents.
出处
《图书情报工作》
CSSCI
北大核心
2007年第6期116-119,126,共5页
Library and Information Service