网页文本的组合分类研究  

Research of Combination Scheme of Web Page Text Classifiers

作　　者：侯小静[1] 舒云星[1] 

机构地区：[1]洛阳工业高等专科学校计算机系,洛阳471003

出　　处：《计算机与数字工程》2007年第6期58-59,共2页Computer & Digital Engineering

摘　　要：在文本分类中,单分器具有某些缺点,如分类模型对样本具有敏感性,并且其分类精度难以提高。本文以朴素贝叶斯为基分类器,利用bagging与boosting技术提出了两种组合分类算法,并将其用于网页文本的自动分类,结果证明组合分类系统提高了分类准确率。There are some shortcoming when it uses single classifier to classify the web page txt, for example, the classification model is sensitive to samples, and the precision is always hard to be improved. There are two algorithm based on bagging and boosting respectively, which take naive bayes as base classifier. We make experiments on web page classification by the two algorithm, and it is proved that the performance of classification system is improved by combination scheme.

关 键 词：分类 装袋 提升 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]