基於知識本體之文獻多維度關聯視覺化導覽平台

Abstract

學術研究必須建立於過去已完成的研究，因此文獻探討非常重要。數位時代下，改變了過去研究者必須東奔西跑的尋找文獻方式，但卻也會造成研究者被許多數位文獻淹沒。若奠基於事先定義好的領域本體論，人類的視覺化資訊處理能力應可協助瞭解文獻中多維度的複雜資料關聯。本研究提出一個文獻導覽平台系統架構，協助使用者依預訂的知識本體導覽領域文獻，並能依循「知識本體與關鍵字」、「引文網路」及「人員關係網路」三種關聯來導覽論文，並與呈現的網路圖形資料互動。最後實作視覺化導覽雛型平台，並以國家圖書館之博、碩士論文網所提供的部分論文為雛形測試資料。

Paper survey is the most important task for conducting research. Nowadays, this work might return huge amount of information owing to digitalized-documentation and high-speed Internet. Based on predefined domain ontology, human visual data-processing capability might help understand multiple dimensions of complicated information relationships. This study proposes a framework for document exploration to assist users to browse domain literature based on predefined ontology. It also guides users to explore the literature based on three different relationships: network of knowledge ontology and keyword, citation network, and people social network. Finally, a prototype was implemented and part of data from Electronic Theses and Dissertations System of National Library were fed to test the feasibility.