根據乏晰觀念網路作乏晰資訊擷取的新方法

Abstract

由於乏晰觀念網路允許使用者藉由觀念與觀念之間的相關性蓷導出 所需求的資訊, 因此以乏晰觀念網路為基礎的乏晰資訊擷取方法已被許多 專家學者所提出. 然而傳統的乏晰觀念網路是以藉於0與1之間的實數值來 表示觀念與觀念之間的相關度, 故其在乏晰資訊擷取的應用上仍不夠彈 性. 在本論文中, 我們將傳統乏晰觀念網路的定義加以擴充, 使得觀念與 觀念之間的相關度可以用區間值或梯型乏晰數來表示. 除此之外, 觀念與 觀念之間的乏晰關係也增加至四種, 即: 乏晰正相關關係, 負乏晰相關關 係, 乏晰一般化關係及乏晰特殊化關係. 在本論文中, 我們更以延伸式乏 晰觀念網路為基礎, 提出了乏晰資訊擷取的新方法, 根據這些新方法所設 計的乏晰資訊擷取系統將更具彈性並更符合使用者的需求. Since the fuzzy concept networks allow us to obtain related information by means of the relevant values between concepts, many fuzzy information retrieval methods based on fuzzy concept networks had been proposed. However,the relevant values between concepts in the traditional fuzzy concept networksare restricted to real values between zero and one. The fuzzy information retrieval methods based on these traditional fuzzy concept networks are not flexible enough in practical applications. In this thesis, we extend the definition of fuzzy concept networks to allow the relevant values between concepts could be interval values or trapezoidal fuzzy numbers. Moreover, four kinds of fuzzy relations between concepts are also provided, i.e. fuzzy positive association, fuzzy negative association, fuzzy generalization, and fuzzy specialization. In this thesis, we also propose fuzzy information retreival methods based on the proposed extended fuzzy concept networks. The proposed methods will make the fuzzy information retrieval systems more flexible to the users.