一种基于特征符号的网页主题信息抽取方法

随着Internet网络的日益普及，Web上的海量数据给文本挖掘尤其是网页主题提取带来了更多的挑战，现有的文本提取方法在保证高准确率的同时无法满足Web挖掘方法的通用性。通过对Web网页结构进行研究，对网页生成树模型进行了改进，找到网页结构的通用规则，提出一种基于特征符号的提取方法CECS（content extraction characteristic symbols），结合相关度对网页主题内容进行提取。实验证明，所提算法具有很高的准确性和通用性。

Content extraction of Web pages based on characteristic symbols

With the popularity of the Internet, the large amounts of data on the Web provides many challenges for data mining techniques, especially for content extraction of Web pages. The existing methods can not guarantee the generality and effectiveness of Web mining approaches. By studying the internal structure of Web pages, this paper proposed an improved document tree model and discovered the general rules for analyzing it. In addition, extracted content from Web pages based on characteristic symbols. The experimental results prove that the proposed method is accurate as well as generic.