以代理人為基礎之個人化知識本體文件搜尋架構

Abstract

網際網路的日趨發達，造成了在網路上搜尋所需資訊越來越艱難。近年來對於這樣的問題，有兩個主要的解決方法，一個是透過資訊檢索的方式來更準確地找到資訊，另一個則是透過個人化的方式來過濾掉資訊。 本研究綜合了以上兩種做法，提出一個新的搜尋架構，並且在搜尋的架構中引進個人化知識本體的概念，透過代理人自動化地從使用者的使用過程、回饋當中，進行學習與分析，個人化使用者的知識本體定義與架構，並在了解到使用者的知識本體後，以概念進行搜尋，進而準確的找詢使用者所喜好的資訊。除此之外本研究提出一個新的思維，透過知識本體概念來群組化使用者，進而從特定領域相類似的使用者中，引進本身所需的知識本體概念代理人，藉以找尋資訊，以加速查詢速度、強化搜尋的準確性。

With the rapid growth of using World Wide Web, the information has become more and more difficult to be located. Presently there are two main solutions to handle this problem. One is to find information more precisely by improving searching algorithm in Information Retrieval; the other is to filter available information by tracking user's behavior. In this paper, we propose a new searching architecture, which combine the two solutions, and we also introduce a concept of 'personal Ontology' into this architecture. By analyzing feedback of user's browsing behavior, agent can realize user's personal ontology, and use the concept of ontology to search, which can find information more correctly and precisely. Besides, we also show a new searching process. By clustering users based on their concepts of ontology, user can find others in the same domain, and use their concept which has been trained properly to find information quickly and more correctly.