黃青淵

Ching-Yuan Huang

碩士論文 (2006)

以概念特徵為基之本體論學習與維護機制研發

Development of a Mechanism for Concept Feature-Based Ontology Learning and Maintenance

關鍵字 Keywords

知識本體, 知識本體維護, 知識管理, 知識本體學習

Knowledge Management, Ontology, Ontology Learning, Ontology Maintenance

摘要

在知識經濟時代，知識管理已成為企業致勝的重要關鍵之一，而能夠促進知識整合、存取、共享及重覆使用的知識本體(Ontology)已逐漸在知識管理領域中扮演重要的角色，因此知識本體（Ontology）有效率的建立、管理與維護是支援知識管理重要關鍵。目前知識本體（Ontology）維護主要的問題之一是如何有效率及正確地學習概念知識進而定位至適當的儲存位置，因此本研究之目的在（1）設計一以概念知識特徵為基之本體論學習架構，包含了多格式輸入、前處理（自然語言處理、概念箤取及特徵標注）、學習與定位與維護功能，（2）依學習架構研發一知識本體學習與維護機制，透過概念特徵標註，運用特徵限制關聯及布林函數進行概念的學習，最後再利用定位圖（Location map）找到正確之定位點，將概念置入知識本體庫中，（3）針對知識本體庫中概念知識的維護，提出知識本體庫維護演算法，包含插入、合併及移除等功能。本研究之成果能有效提昇本體論學習與維護之效率及正確性，進而實現知識管理與分享之目標。

Abstract

In the knowledge economy Age, Knowledge management has become important key factors of enterprise success. ontology act as vital support role of knowledge management for facilitated knowledge integration、access、share and reuse. So building、management and maintenance ontology efficiently are importance of supporting knowledge management. The one of main problem of ontology are efficiently and accurately to learning and location to suitable store address. For this reason, In this thesis objective are (1) Designing a frameworks for Concept Feature-Based Ontology Learning, including Import of diversity entity knowledge、preprocess（Natural Language Process、Concepts extraction and Feature Annotation）、Learning and Location、Maintenance and User interface. (2) We according to framework to development of an Mechanism for Concept Feature-Based Ontology Learning and Maintenance through annotate out concept feature and feature constraint relations for concept learning. Finally, Find out Concept Location on the Is-A hierarchy from Location map. (3) We take a function of ontology maintenance Algorithms , including Insert、Merge and Delete. In our these result to be able to promotion efficiency and accuracy of ontology learning and maintenance. And then to carry out an objective of knowledge management and share.