語意擴充於本體論相似度組合之研究

Semantic Enrichment in Similarity Combination for Ontology Mapping

黃冠豪(Kuan-Hao Huang)

指導教授 ： 劉立頌

國立中正大學/工學院/電機工程學系/碩士(2014年)

摘要

本體論為知識分享的工具，不同領域間知識的傳達需要依靠本體論作為溝通
的橋樑，透過本體論能讓不同領域間知識的傳遞互通有無，各領域的專家學者存
在不同的專業背景，各領域的專家擁有著屬於該領域表達概念的方法，在這樣百
花齊放的表達方式之中，即便是在相同領域中個人的表達方式亦可能造成格式、
內容等差異，每個領域間始終存在著如何建立溝通管道的問題；本體論對映是透
過在不同的本體論間建立連結，讓使用者能夠清楚且方便的了解領域的內容，使
得不同形式的知識能夠在該領域內互通有無。本體論對映的核心基礎，在於如何
找出兩者間關聯，因此透過相似度的計算找出兩者間關聯程度，然而早期的研究
團隊所採用的處理方法，大多採用單一相似度的計算，然而隨著本體論規模的擴
大，複雜程度亦隨之提升，因此便頇要透過各種不同的角度來進行相似度處理，
使各種陎向的相似度均能獲得妥善的處理，並且透過相似度的組合將之結合，使
兩本體論間概念做出連結；由於在本體論內部常常需要清楚的定義進行處理，在
語意內容的相似程度方陎，往往僅處理部分的相同、異詞性資訊，對於內部的屬
性關係並未多加著墨，因此希望透過擴充語意內部屬性的關聯，增加本體論搜尋
的範圍，並進行多種相似度的組合，使系統能夠挑選出適當的配對組合推薦給使
用者，以增加整體系統的評估能力。

關鍵字

語意擴充 ； 本體論對映 ； 本體論 ； 相似度組合

並列摘要

Ontology can be seen as a knowledge sharing tool to communicate between
different fields of knowledge. Each expert has their own way to design their
ontologies using their own structures. This introduces a challenging problem of how
to establish communication channels between these diversities. Ontology mapping is
a good method to solve these problems by building relations between these ontologies.
The basic core of ontology mapping is to compute the similarities between these
concept pairs. In early years, researchers considered only single similarity for their
ontology mapping. Recently,the scale of ontology has become more complicated.
More researchers try to filter out all the possible similarities to combine concepts
together to get the better result of mapping. One of the important similarity is the
semantic similarity. Some of the researchers use just only synonyms or antonym to
find out the relation. In semantic similarity, how to get the attributes or properties to
improve the scope of searching is an important part. This thesis presents a method to
improve the system by enriching the semantic similarity to expand the scope of the
searching area. Constituting semantic similarity with other relations make the system
have a better result in finding the correct similarity combinations and produce high
possibility of successful mapping.

並列關鍵字

Similarity Combination ； Semantic Enrichment ； Ontology Mapping ； Ontology

參考文獻

[1] W.N. Borst, Construction of engineering ontologies for knowledge sharing and

and comparative review,” AI magazine, vol. 18, no. 3, pp. 53-74, 1997.

[3] T.R. Gruber, “A translation approach to portable ontology specifications,” in

[4] R. Studer, V.R. Benjamins, and D. Fensel, “Knowledge engineering: principles

First International Conference of FOIS'98, vol. 46, IOS press, pp. 3-15, June