摘要
1.引言随着Internet的迅速发展,网上信息成几何级数不断增长,如何从中找出人们需要的文献是信息检索要处理的重要问题.
A novel pattern classification method, called nearest feature line(NFL) has been proposed . In this paper we propose a novel classification method, called the k nearest feature line (kNFL),for text classification and approach it to application- The experiment result shows that kNFL presents better performance than NFL, k nearest neighbors (kNN) and nearest neighbor(NN) ,especially when the training sets are small.
出处
《计算机科学》
CSCD
北大核心
2002年第1期47-48,56,共3页
Computer Science