Knowledge categorization is the shared recognition of people to a certain knowledge domain. It serves the purposes such as description, interpretation, communication, and inference of a domain of knowledge. Because the application of knowledge in practice for problem-solving can involve multiple domains, therefore how to construct logic relationships among the knowledge categorizations is often the key to problem-solving. This study takes the case of outpatients’ perception of signs and symptoms as an example domain to explore how knowledge categorization and integration could help outpatients query and choose from the departments of a healthcare institute at registration. This study uses ontology modeling to develop the needed knowledge categorization and problem-solving models. The major contents include: (1) Clarifying the content of the knowledge sources, including the knowledge categories of symptoms and illness; (2) Constructing a domain ontology of general knowledge sources consisting of common conceptual structure and instances to provide reference standards or terminology when communicating with other domains; (3) Establishing an objective-oriented task-ontology by developing the relationships and logic among the knowledge sources in accordance with the needs of problem-solving, and then collecting existing facts as instance knowledge in accordance with the knowledge schema of the concepts; and (4) Developing a set of inferable semantic rules for problem-solving to infer the implicit knowledge based on the aforementioned factual knowledge. The experiment results show that the procedures of knowledge categorization and integration developed in this study, with the modeling of domain ontology, task ontology, and inference rules, have preliminarily achieved the purpose of solving the problem of matching outpatients’ signs and symptoms with the suitable hospital department. Furthermore, the results of this study have simplified the future maintenance and expansion of the domain content knowledge and thus enabled effective knowledge integration. pp. 69-89知識分類是人類對領域的共同認知，利用結構化方式建立系統性表達，以提供描述、解釋、溝通、及推論。由於實務上的應用問題通常涉及多種領域，因此如何建立它們的知識邏輯是解題關鍵。本研究以個人對體徵症狀（signs & symptoms）的感知為例，探討藉由知識分類及整合，最終能「查詢」特定醫療院所的就診參考。本研究利用知識本體（ontology）技術，發展所需的知識分類與解題模型，主要的內容包括：（1）釐清此議題的知識源內涵，例如徵狀、疾病等知識分類；（2）將一般化的知識源建置為「領域本體」，其內容是由共通性的概念架構及實例共組而成，以利提供其他領域在溝通時做為參考標準或術語；（3）以解題需要來發展各知識源之間的關聯與邏輯，建立目標導向的「任務本體」，再依據各概念的知識框架（schema），收集現況事實為實例知識；（4）最後，發展可推論解題的「語意規則」，並以前述的事實知識為基礎，推導隱含性知識。由實驗結果顯示：本研究發展的知識模型整合程序，強調領域本體、任務本體、及推論規則的模型設計，已初步達到解決以體徵症狀查詢就診科別，也簡化後續知識之維護及擴充，達到知識整合之效用。頁64-83