计算机工程与应用 ›› 2017, Vol. 53 ›› Issue (18): 236-242.DOI: 10.3778/j.issn.1002-8331.1604-0039

• 工程与应用 • 上一篇    下一篇

基于WSFT模型的深层网文本获取方法

杨贯中,李虹萱   

  1. 湖南大学 信息科学与工程学院,长沙 410082
  • 出版日期:2017-09-15 发布日期:2017-09-29

Approach based on WSFT for crawling deep web

YANG Guanzhong, LI Hongxuan   

  1. School of Information Science and Engineering, Hunan University, Changsha 410082, China
  • Online:2017-09-15 Published:2017-09-29

摘要: Ajax技术在深层网(Deep Web)网站开发中得到了广泛应用。针对Ajax页面多状态、状态之间强关联的特性,提出一种构建WSFT(带权状态融合树)模型的方法,来进行Ajax页面文本信息预处理。引入了文本特征树作为状态指纹进行状态捕获,优化了当前Ajax页面数据采集方法,同时通过StatusRank方法计算状态转移权值来分析状态迁移信息,最后生成WSFT。实验证明,该方法能有效地获取Ajax页面多状态文本信息,并且有助于后续Web挖掘的重要文本内容提取。

关键词: Ajax爬虫, 带权状态融合树, 文本挖掘, 文本特征树

Abstract: Ajax technology has been widely applied in deep web application development. This paper constructs a Weighted State Fusion Tree (WSFT) model to pre-process the text information in web page with Ajax technology which has multiple states with strong correlation. Firstly, the current approach of Ajax page data collection is optimized by regarding text feature tree as a fingerprint to traverse through the multiple states. Secondly, the transition weight with StatusRank method is calculated for each states of the Ajax page. The state transition information is analyzed. Finally, a WSFT is generated. The experimental results show that the proposed method can effectively obtain the text information in Ajax page with multiple states, and help the follow-up important text extraction of web mining.

Key words: Ajax crawler, weighted state fusion tree, text mining, text feature tree