关键词: Web信息抽取, 林产品贸易语义词典, 语义信息熵, 模板, 目标信息定位

Abstract: Based on the shortages of the existing Web information extraction technique in the presence of the accuracy is not high, a low degree of automation and the weaker commonality, combined with the structured storage needs of information source in forest products trade Web information push, a new algorithm on forest products trading Web messages structuring based on semantic is proposed. The paper analyzes and takes advantage of forest products trade Web information feature, and combined with the basic principle of semantic recognition, it constructs of the forest product trade semantic dictionary, uses the layout features of the target information that need to extract in the Web pages at the same time and combined with the information entropy theory, a method of target information automatic extraction based on the semantic information entropy is proposed to extract target information, and the information is stored in the database as a structured form. The experiments on actual forest product trade Web pages information extraction, prove that this algorithm can reduce manual intervention and has good value in processing information source in forest products trade information push.

Key words: Web information extraction, forest product trade semantic dictionary, semantic information entropy, template, target information location