机构地区: 北京科技大学计算机与通信工程学院
出 处: 《计算机科学》 2008年第12期167-170,共4页
摘 要: 文本分类是数据挖掘的一种应用,分类器的设计是其中最重要的一个环节。与那些基于统计方法的分类算法比较,给予规则的分类算法在很多情况下更具优越性。提出了一种基于粒运算的方法,通过建立粒网络生成分类规则,从而实现文本分类的方法。本算法通过从最大的粒中提取较小粒,直至产生最小粒的过程建立起一个粒网络,从而产生分类规则,实现文本的分类。 Text classification is one of the practices of knowledge discovery and designation of the classifier is the most important part. Comparing with the methods based on statistic theory, classification based on rule learning is a better one in some situations. A granular computing approach was proposed to learn classification rules by constructing a granu- le network. The algorithm is involved in a refining process from the largest granule to the smallest one. By constructing a granule network, text classification rules are learned.