透過您的圖書館登入
IP:54.196.52.45
  • 學位論文

發展使用者認知為基礎的知識本體應用於資訊檢索系統

Developing Consumer Cognition-Aware Ontology for Information Retrieval System

指導教授 : 戚玉樑

摘要


為改善過去詮釋資料僅對資料作描述,及未定義結構化資訊與其領域正規知識表達的映對關係,本研究使用語意的註記方式,以XML-based的註標語言,例如:RDF與OWL來表達資料間更豐富的關係,將資料視為概念,利用知識本體技術對資料背後的領域知識塑模,對概念下定義,以推論出資料間的隱含關係。知識本體技術已成為取代詮釋資料的方法之一,但是目前的知識本體設計,仍以專家的角度來建置,使得系統不理解使用者的語意,也無法滿足其需求。本研究在探討以使用者需求為出發點的知識本體工程建置,考慮使用者與專家對於專門領域的認知差異,因為使用者對於不熟悉的領域,會擁有不足的專業理性知識,所以會以感性想法來理解領域的內容。除領域知識本體外,發展認知知識本體架構,令所建置的知識本體庫不僅具備此專門領域的知識外,更滿足使用者的內隱需求,即包含其隱含的感性知識,使系統理解使用者的語意,檢索系統提供更多元以及人性化的查詢。從一開始認知蒐集階段,找出使用者所具備的「知」的程度,納入使用者所擁有的直覺、知覺、情緒等,作為知識本體設計的考量,利用感性工學(Kansei Engineering)的方法,透過問卷擷取使用者對領域中標的物的認知,找尋標的物特徵與使用者認知之間的關係,以設計認知知識本體模型。對於領域中標的物的描述,從單純的資料層級描述提升為語意層級的描述,系統的開發也從開發者角度轉換為使用者的角度,最後,以精確率與召回率評估此檢索系統,發現此系統能滿足不同專業程度的使用者,並縮小相關資訊的範圍,提供使用者一個符合不同程度需求的人性化智慧檢索系統。

並列摘要


This study develops a cognition-aware ontology in addressing poor emotional representation. Cognition-aware design refers to define the subjective emotional relationships between ordinary users and knowledge bases. The primary task of this study is translating human emotions into sensory-based properties that ontology intends to communicate through affective descriptions of their physical attributes. Since ontology consists of concepts and their relationships, the cognition-aware design must be applied to the design elements of concepts. To implement cognition-aware of ontology development, finding the elements of concepts such as physical attributes and their corresponding cognition properties are essential. Several methods have been utilized particularly in the expertise acquisition stage such as a questionnaire test, an affinity diagram, and statistical analysis. A walkthrough example of identifying dog breed is given to demonstrate the benefits of using a cognition-aware approach in the ontology building processes, including knowledge acquisition, concepts modeling, and ontology representation. Empirical lesson indicated that cognition-aware design is necessary especially when the users are non-experts. Experimental results indicated the new design helps users obtain appropriate results after inputting affective descriptions. The performance evaluation is positive.

參考文獻


[7]戚玉樑,「知識擷取與知識表達協同程序於建構知識本體的概念架構」,資訊管理學報,13(2),2006年,頁193-215。
[7] Choi, B., Henson, K., Raghu, T.S. and Vinze, A., “Knowledge Sharing Ontology
[1] Alberts, L. K., “YMIR: an Ontology for Engineering Design,” Ph. D.
Dissertation, University of Twente, 1993.
evaluation of ontology languages for geographic information retrieval on the

被引用紀錄


吳俊逸(2011)。以專利本體論工程為基之智財服務防禦式決策支援系統〔博士論文,國立清華大學〕。華藝線上圖書館。https://doi.org/10.6843%2fNTHU.2011.00618

延伸閱讀