透過您的圖書館登入
IP:3.134.104.173
  • 學位論文

建構一個半自動化Semantic Web的RDF/XML產生與展示系統

Building a semi-automatic Semantic Web based RDF/XML generation and presentation system

指導教授 : 陸承志
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


由於網際網路的普及與社會的變遷,利用網路查找資訊已是非常普遍。不過,現代人在茫茫網海中搜尋的資訊量不但過多,還必須由檢索者親自耗費許多時間過濾搜尋結果;因此,Tim Berner's Lee 提出「Semantic Web」概念,希望將Web內容加上語意結構,以機器可以處理的方式來定義與連結網頁資料,使得資料得以自我解釋,方便被索引與搜尋。 目前建構Semantic Web的方法是使用資源描述框架(RDF,Resource Description Framework)以XML格式來儲存語意資訊,不過RDF所能表達的資料較為原始並且缺乏大量產生RDF/XML的工具;此外,使用者無法利用專家對特定領域的知識與經驗來改善搜尋精準度,使得Semantic Web理想遲遲無法普及。 為解決上述困境,本研究提出一分階段的「擷取、分析、展示」架構,利用專家協助建構的Ontology來半自動產生包含語意的RDF/XML文件,並實作一個系統來驗證本架構的可行性:首先在專家協助下建構目標領域的Ontology,接著由輸入資料中擷取出語意資料,並先以本研究提出的Pseudo RDF格式儲存,然後再轉換成符合標準語法的RDF/XML,最後以網頁介面展示系統建構所得。經實驗成果顯示,本研究架構不僅可行,也可有效加速產生符合精確語意的Semantic Web。

關鍵字

Semantic Web RDF Ontology Dublin Core XML Pseudo-RDF Metadata

並列摘要


Using search engine to find information is common to all internet users. However, most search engines return too many results than what users want. Users then are required to spend a lot of efforts to filter out what they really want. The Semantic Web concept, proposed by Tim Berner’s Lee, is intented to add semantic description to web contents so that they can be easily indexed and searched by machines. RDF is a tool used to describe Semantic Web contents. Writing RDF on web contents is not only tedious, but also is hard to leverage expert’s experience on special domains. This becomes an obstacle to wide adoption of Semantic Web. A three staged “extraction, analysis and presentation” framework is proposed in this research to remedy the problem stated above. This framework is intended to semi-automatically generate RDF/XML for web contents with ontology of a special domain. We build a prototype system to validate this framework. First, we build ontology of Cihu poetry published at Sung Dynasty with help from domain experts. Then we extract semantic data from some sample poetry into Pseudo-RDF form. Next, we transform Pseudo-RDF into standard-compatible RDF/XML and provide a web interface to demonstrate the results in graphic form. Our results show that generator can be easily to re-generate Semantic RDF/XML once either the domain ontology or web contents are changed.

並列關鍵字

Semantic Web RDF Ontology Dublin Core XML Pseudo-RDF Metadata

參考文獻


〔24〕 Michael Denny ,”Ontology tool survey” , http://xml.coverpages.org/Denny-OntologyEditorSurveyTable20021111.html
〔32〕 吳政叡,「從都柏林核心集看未來資料描述格式的發展趨勢」,圖書館學刊 26 期,(民 86 年 6 月),頁 11-18,網址http://dimes.lins.fju.edu.tw/pub/fju-lins-26/dublin1.htm
〔1〕 Google Search Engine , http://www.google.com
〔2〕 Open Directory Project , http://www.dmoz.org
〔3〕 Sergey Brin, Lawrence Page , ”The Anatomy of a Large-Scale Hypertextual Web Search Engine” , Computer Networks and ISDN Systems , volume 30 no 1-7 , pp. 107-117 , 1998

延伸閱讀