基于dom和网页模板的信息抽取-计算机应用技术专业论文.docxVIP

中山大学硕士学位论文 中山大学硕士学位论文 基于DOM和网页模板的信息抽取 Information Extraction Based on DOM and Web Template Computer Application and Technology Name：Su Wenjian Supervisor：Chen Youqing Abstract With rapid development of Intemet，Intemet has been a huge knowledge repository．In order to make 900d use of infc}rmation OFI the Interact，infonnation extraction comes out．Information extraction belongs to the branch of artificial intelligence．People can extract interesting information from web pages by means of it． The paper first introduces the history and the research at home and abroad of information extraction．Then it compares information extraction to several relevant technologies．Atier that．it makes a brief analysis of information extraction and introduces the key task，the main way of research，the system and the evaluation standard of information extraction． Information extraction studied by the paper iS based on DOM(Document Object Model)and web template．DOM iS a standard of W3C．The structure of web pages Can be well described with it．According to the definition ofDOM．the paper describes the structure ofweb pages by constructing HTML parsing tree． Because many web pages on the intemet are built with web templates，the paper proposes a new way of inducting web pages by reference of the achievement of the former．The way of inducting web page templates in the paper is for the web pages that are 1aid out by table element．By inducting web template，the noise information in web pages Can be eliminated SO that it can raise the precise rate of information extraction． Information extraction in the paper iS based on the path of DOM．The extraction rule iS based on relative path instead of abSolute path．The extraction rule based on relative path can reduce the impact of the variance of the structure of web pages． Besides，extraction rule based on relative path is self-adaptive．When the structures of web pages are completely different，extraction rule Can be rebuilt automatically． The paper develops