一套將關聯式資料庫對映至知識本體的方法

An Approach to Mapping Relational Databases to Ontologies

鄭伊善(Yi-Shan Cheng)

指導教授 ： 蔡益坤

國立臺灣大學/管理學院/資訊管理學研究所/碩士(2007年)

https://doi.org/10.6342/NTU.2007.02437

摘要

網路服務技術允許應用程式透過網路標準以一致的方式被存取，使得它們就像是不同平台上的軟體元件。隨著網路服務的規模越來越大，找到適合的網路服務成為一件重要的工作。現今大多數使用者使用關鍵字搜尋以及人工閱讀服務的內容描述這種方式來尋找網路服務，這件工作相當的費時且沒有效率。語意網是一種利用知識本體加以註釋並且機器可以理解的網路，透過將語意網應用到網路服務上，我們可以使得網路服務可以被機器理解。大多數人使用知識本體定義來當作語意網上語意註釋的基礎，因此從現有的資訊中自動化的產生知識本體事件意義重大的工作。關聯式資料庫是最為普遍的資料儲存系統，對於那些想要參與語意網路服務領域的服務提供者來說，找到他們使用已久的關聯式資料庫與知識本體間的對應關係是一個關鍵性的議題。
　　在這篇論文中，我們開發出一套方法可將一個關聯式資料庫綱目對映到某特定領域中一個已有的知識本體，這個使用已久的關聯式資料庫藉此可以被語意網路服務所運用。第一階段是元素層級的媒合，目標是衡量單一一個表格和單一一個類別的相似度。這個階段的配對將所有單一一個表格和單一一個類別可能的組合搭配出來後，使用自然語言處理技術藉由它們的名字以及它們的欄位和屬性的名字來計算相似度。第二階段是綱目層級的媒合，考慮的是這兩種資料綱目的整體結構。直覺地，對一個表格和一個類別來說，他們相關的表格和相關的類別之間越接近，他們會實際被對映在一起的機率就越高。綱目層級的媒合者由它們的子類別/表格、超類別/表格、參考和被參考到此表格的那些表格、以及那些透過某些物件屬性與此類別產生關聯的類別來測量結構上的相似度。一個表格和一個類別整體的媒合分數就是由這兩個媒合階段計算出來的媒合分數所組成的。由於當初開發這兩種綱目的目的不同，加上各種不同的設計風格，期待高準確率的完全自動化對映是不切實際的。因此對於每一個表格，我們的系統會選擇最高分的幾組對映方式當作候選來讓使用者選擇。

關鍵字

知識本體 ； 知識本體語言 ； 語意網 ； 對映 ； 網路服務 ； 關聯式資料庫

並列摘要

Web Service technology allows uniform access via Web standards as software components residing on various platforms. As the scale of Web Services becomes larger, finding the suitable Web Services is a non-trivial work. Currently, most users find Web Services by keyword search and manually read the description contents of the services. This work is time consuming and lack of efficiency. The Semantic Web is a machine understandable Web annotated by ontologies. By applying Semantic Web technology to Web Services, we can make Web Services to be machine-understandable. Most people use ontology definitions as the basis of semantic annotation on Semantic Web. Consequently, automatic construction of ontologies from existing information is relevant. Relational databases are the most common type of data storing system. For service providers who want to participate in Semantic Web services, to find out the mappings between their legacy relational database to ontologies is seen as an critical issue.
　In this thesis, we develop an approach to mapping a relational database schema to a specified domain ontology herewith the legacy relational database could be applicable
to Semantic Web services. The mapping algorithm consists of two phases to find the optimal mappings. The first phase is the Element-Level Matching which aims to measure
the similarity of a single table and a single class. This phase of matching matches all possible combinations of a table and a class, measures their similarity by their names and the names of their columns and properties, using natural language processing techniques. The second matching phase is the Schema-Level Matching, which considers the overall structures of both schemas. Intuitively, for a table and a class, the more their associated tables and associated classes are similar, the higher possibility that they would be actually mapped. The matchers of the Schema-Level Matching phase measure the structural similarity by their sub-tables/classes, super-tables/classes, the tables that referring to/referred to by the table, and the classes associated by object properties. The overall match score of a table and a class is composed of the two match scores which are calculated
by the two matching phases. Because of the different development goal of these two schemas and the various designing styles, it is impractical to expect the high accuracy of full-automatic mapping. Our system therefore would choose several top mappings for each table as candidates to be chosen by the user.

並列關鍵字

Ontology ； OWL ； SemanticWeb ； Mapping ； Web Services ； Relational Databases

參考文獻

[1] I. Astrova. Reverse engineering of relational databases to ontologies. Proceedings of 1st European Semantic Web Symposium (ESWS), Heraklion, Crete, Greece, LNCS, 3053:327–341, 2004.

[2] T. Belwood et al. UDDI Version 3.0.

[7] R.H.L. Chiang, T.M. Barron, and V.C. Storey. Reverse engineering of relational databases: extraction of an EER model from a relational database. Data & Knowledge Engineering, 12(2):107–142, 1994.

[8] E. Christensen, F. Curbera, G. Meredith, and S. Weerawarana. Web Services Description Language (WSDL) 1.1, 2001.

[9] E.F. Codd. A relational model of data for large shared data banks. Communications of the ACM, 13(6):377–387, 1970.