基于极限学习机的网页分类应用  

Classification of web pages based on extreme learning machine

作　　者：陈先福[1] 李石君[1] 曾慧[1] 

机构地区：[1]武汉大学计算机学院,武汉430072

出　　处：《计算机工程与应用》2015年第5期102-106,共5页Computer Engineering and Applications

基　　金：国家自然科学基金(No.61272109)

摘　　要：极限学习机ELM不同于传统的神经网络学习算法(如BP算法),是一种高效的单隐层前馈神经网络(SLFNs)学习算法。将极限学习机引入到中文网页分类任务中。对中文网页进行预处理,提取其特性信息,从而形成网页特征树,产生定长编码作为极限学习机的输入数据。实验结果表明该方法能够有效地分类网页。ELM extreme learning machine is different from traditional neural network learning algorithm（such as BP algorithm）, is a highly efficient Single hidden Layer Feedforward Neural network（SLFNs）learning algorithm. In this paper,ELM is introduced to Chinese web page classification task. Trait tree of web page is formed after pre-processing the Chinese web and extracting its characteristic information. Fixed-length coding is produced and took as input data of ELM.Experimental results show that the method can effectively classify web pages.

关 键 词：极限学习机 中文网页分类 神经网络 网页特征提取 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]