本文主要以「健保藥品給付規定」做為準則,建置知識庫獨立於推論引擎的決策支援系統,使得知識庫可重複使用,且修改時不需更動推論引擎的程式碼。知識庫主要是應用Protégé2000為儲存工具,並建構健保藥品給付領域本體論(Domain Ontology)用以描述疾病與藥品概念類別與關聯,準則模式化的方法為應用方法本體論(Method Ontology)以表示健保藥品給付規定的準則。我們將給付規定細分成基本規則(rules),再加以組成健保藥品給付規定的準則。由於為使用本體論階層式的設計方法,故可以提昇準則再利用性及領域知識的共享性,提供給該領域專家及相關知識工程師使用。此外,本文更提出本體論知識庫和醫療院所資料庫整合機制,以增加知識庫再使用的彈性及減少使用上的限制。知識庫及本體論的建置平台皆使用Protégé2000,並以其提供的介面來呈現、組合和管理健保藥品給付領域本體論和方法本體論知識庫,推論引擎實作部份,以使用跨平台的Java語言。
This thesis is to build a decision support system based on the Drugs guideline for Dyslipidemia published by the Bureau of National Health Insurance.We make use of Protégé 2000 to construct domain and method ontology knowledge bases that are separated from the inference engine of the system. Therefore, The knowledge base can be reused and modified without modification of the program of the inference engine.The domain ontology is used for representation of the domain knowledge by using classes and slots. The classes represent the disease and drugs groups, and the slots represent the attributes and the relationships between the classes. The method ontology is used for representation of the drug guideline. The guideline is divided into primitive rules, which in turns are used to compose a more complicate rule, called the composite rule. The drug guideline is then composed by a set of primitive and (or) composite rules. Thus, the rules can be reused and shared easily. Because the ontological domain and method knowledge bases are represented, managed and displayed by Protégé 2000, they are easily understood by the domain experts, and hence they can be more likely shared and reused.The decision support system is implemented by java code.
為了持續優化網站功能與使用者體驗,本網站將Cookies分析技術用於網站營運、分析和個人化服務之目的。
若您繼續瀏覽本網站,即表示您同意本網站使用Cookies。