在一般的搜尋引擎使用經驗上，使用者往往會搜尋到大量的資訊，其中更可能包含了與使用者原始目的不相同的文件。主要原因是因為電腦在搜尋文件的過程中往往只是針對字面上的詞去做比對，而沒有詳細去分析該字詞的真正語意為何。 本研究的目標為提升搜尋引擎的效率並協助使用者搜尋正確的文件，以強化對於搜尋結果的滿意度以及實用度。我們利用了本體論方法來強化搜尋階段所需的數位內容的語意描述。針對實驗室數位圖書館的環境下，我們提出了數位典藏本體論，用以描述文獻內容的意義，領域知識，以及使用者描述資訊。搭配邏輯推理相關機制，發展了推論式的智慧型搜尋引擎。其目的為協助使用者搜尋學術相關文獻，並可根據使用者的背景知識提供其容易理解的相關資訊。 實驗結果顯示推論式搜尋方法可以有效的提升搜尋引擎的效益。有別於傳統關鍵字搜尋的機制，本研究的成果可以有效的推演出使用者的需求並精確的搜尋出符合使用者語意的文獻內容。 According to the experience of retrieving content via search engine, people often get a great deal of responded information. Somehow, part of them is irrelevant to user’s intention. The main reason caused this problem is the lacking of enough semantic description for digital content during the analyzing, searching and matching processes. The purpose of our study is to improve the searching efficiency, user’s satisfactions and practicability. We apply ontology theory to content description, which is used in matching process. In digital library application, we proposed Digital Library Ontology to establish the descriptions of contents, domain knowledge and user profiles. Utilizing reasoning technique, we developed an inference-based intelligent search engine to assist literatures retrieval based on users' background knowledge. Experiments show the proposed intelligent search engine can efficiently improve searching performance. To distinguish from traditional keywords search, our approach can provide better searching result, which is based on deduced needs from user. The responded literatures are verified to have better comprehension to user.
