以本體技術為基礎的知識庫建置程序及其應用

 

The Ontological Techniques for Building Knowledge Bases: Its Procedures and Applications

 

戚玉樑

中原大學資訊管理研究所

桃園縣中壢市(32023)中北路200

maxchi@mis.cycu.edu.tw

 

摘要

本研究探討如何藉本體技術(Ontological techniques)來建置知識庫,在本體知識庫的發展上,我們建議須分別依照規劃、設計、測試修正、佈署及整合擴展等五項階段進行,由於本體知識庫主要是由概念架構及其關係所構成,因此設計階段是其成效的重要關鍵。為收集對真實世界的認知,我們利用正規概念分析法來獲得所須的概念及其關係,並發展常用的邏輯模式,以降低知識表達的轉換障礙。本研究利用上述發展程序並結合Ontology開發工具,以植物學門的維管束植物為例,進行實際應用上的驗證。由結果顯示,本研究所提出的發展程序,特別是設計階段採用的方法,確可協助建置者加速本體知識庫的建立。

 

關鍵詞:本體論、知識庫、描述邏輯、正規概念分析法、推論。

 

 

Abstract

For ontology builders, creating suitable conceptual structures of specific domain is increasingly problematic. Though humans have been building ontologies for years to express subject area of interest, an efficient ontology building process is very much ad-hoc. This paper describes development procedures that comprise planning, design, testing & modification, deployment, and integration. This paper particularly emphasizes the design procedure that involves utilizing Formal Concept Analysis (FCA) to gain conceptual structures and Description Logic (DL) to represent concept relationships in expressions. The paper provides a walkthrough example that utilizes Vascular plants as an applied domain. The empirical findings indicate that development procedures have advantaged in guiding ontology builders to create ontological knowledge bases step by step. Moreover, the cognition extraction is useful and connectable for builders and other tools.

 

Keywords: Ontology, Knowledge base, Description logic, Formal concept analysis, and Inference.