以主題地圖建構索引典之語意網路模型

Constructing a Semantic Network Model of Thesaurus: Topic Maps Approach

主題地圖是一種分散式知識表示法，可以有效的整合網路上的知識庫，從而建構出一個分散式知識管理系統。本文首先從主題地圖之發展切入，介紹其相關標準及其內涵；接著探討知識組織、書後索引、術語彙編與索引典等相關概念；然後實際以教育論文線上資料庫的中文教育類詞庫中，選擇十九個主題詞及其替代詞作為實驗用素材，歸納出其詞間關係並建構語意網路模型，再將此語意網路依據國際標準組織的主題地圖規範進行實作，並在伺服器端建置主題地圖處理器，成功的完成一個實驗性質的線上索引典；最後，我們從主題地圖之相關應用議題與未來發展探討其潛力與限制。

Topic map is a distributed method of knowledge representation; it can integrate the knowledge on the network and build a distributed knowledge management system. This paper begins with a survey of the standards and content of topic maps and follows by concept definitions of knowledge organization, index, glossary and thesaurus. Then, an experiment is run to construct a semantic network model by selecting 19 descriptors and their non-descriptors from the thesaurus. An experimental on-line thesaurus is completed with topic map processor successfully installed on the server. The paper ends with a discussion on the potentiality and limitation of topic maps.