K特征线法在文本分类上的应用  

The Application of K Nearest Feature Line Method to the Text Classification

作　　者：杨昂 

机构地区：[1]湖南大学计算机系,长沙410082

出　　处：《计算机科学》2002年第1期47-48,56,共3页Computer Science

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.

关 键 词：文本信息处理 K特征线法 机器学习 模式识别 文本分类 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TP181[自动化与计算机技术—计算机科学与技术]