近年來,語意式搜尋(Semantic Search)儼然已成為語意網(Semantic Web)最炙手可熱的研究。而分類方法在搜尋的處理中扮演一個非常關鍵的角色,因為它在搜尋初期就能消除語意不恰當的資訊。然而,許多與 Folksonomy 相關的應用卻都遭遇一些阻礙。所幸的是現存的語意技術–邏輯描述(Description Logic)似乎可以解決這方面的問題。另一方面,如何有效地利用網際網路上的異質性本體知識,以達領域知識的完整性,亦是我們的研究目標。因此,本研究根據面向分類法(Faceted Classification),並以法則(Rule)取代其標記法,萃取一面向(Facet)的邏輯層,以組織現存可用的本體。我們先經由全盤地掌控類別之間的階層以及 parts-whole 關係,以強化類別架構。接著,根據我們提出的一系列的步驟,萃取面向邏輯層。實作的結果證明我們的方法能夠有效地整合不一致甚至異質性的本體。此外,本研究也歸納出挑選面向應遵守的原則。我們所實作的面向瀏覽器也完全符合這些原則,而得到較佳的語意式搜尋結果。
Semantic Search has become an active research of Semantic Web in the recent years. The methodology of classification plays a pretty critical role in search process, because it can disambiguate irrelevant information in the beginning of process period. However, the applications related to Folksonomy suffer from much obstruction. Fortunately, It is possible to utilize an existing semantic technologyDescription Logic to eliminate the problems resulted from Folksonomy. On the other hand, how to effectively utilize heterogeneous ontologies over the Internet to gain the integrity of domain knowledge is also our research objective. In this thesis a faceted logic layer is abstracted in order to organize existing available ontologies according to the methodology of Faceted Classification while using rules to replace its notation. First of all, the category framework is strengthened by thoroughly control hierarchical or parts-whole relationships between categories. Following that, a faceted logic layer is abstracted according to a series of steps we proposed. The implementation result demonstrated that our approach can facilitate the integration of inconsistent or even heterogeneous ontologies. Additionally, this thesis also generalizes the principles of picking appropriate facets. The developed facet browser completely complies with these principles so that better semantic search result can be obtained.