關聯圖形化之健康網站資訊呈現技術應用：以SARS網站資訊查詢與呈現為例

Applying the Graphical Relational Representation Technique to Represent Healthcare Web Information: Taking the Query and Representation of SARS Web Information as an Example

當民眾擔心傳染病爆發時，網路常是他們獲取相關資訊的來源之一。但研究發現民眾並不滿意目前網路上協助尋找資訊的工具。因此，如何改善資料蒐集整理及網站資訊呈現方面的問題，一直是重要的研究課題。為了改進HTML對於目前網頁文件資訊呈現方式的限制，本研究採用XML做為文件資料庫的建置及交換格式。先由網站文件管理者建立文件的相關屬性，再透過系統搜尋處理建立文件之間的關聯性，然後以SVG格式繪圖來呈現使用者瀏覽介面中的關聯圖形並表現更多的互動功能。以此改變目前網站提供的資訊瀏覽方式，讓進入系統網站內的一般網路資訊搜尋者可以藉由得知文件之間的相關性以及直覺化的圖形點選方式，來獲取所需的資訊。鑑於2003年嚴重急性呼吸道症候群(SARS)爆發流行時，網路上龐大且混亂的資訊曾造成民眾相當大的困擾，故本研究以SARS的網路資訊作為系統開發與評估的案例。結果顯示，對於本系統所提供的操作瀏覽資訊方式，使用者覺得容易使用且有助於瞭解資訊的內容及相關性，因此可改善瀏覽搜尋資訊的效率。在未來應用上，建議可以結合關鍵字搜尋引擎或是自然語言處理系統，以建置整合性更高的網路資訊。

The web site has been one of the main information sources when people are concerned about the outbreak of the infectious diseases. However, one previous study showed that consumers are not satisfied with current web information search tools. Therefore, how to improve data collection and representation of web information has been a major research issue. To resolve the restrictions of the HTML-based content representation approach, the XML format was used to design the documentation base for information storage and exchange in this study. The website manager was expected to preset the attributes of all the pertinent web articles such that the relations among articles could be established. Those relations could then be presented by an interactive graphic interface based on the SVG language for users to query the information intuitively. The website information about severe acute respiratory syndrome (SARS), of which the enormous amount of unmanaged information had caused great confusion to users during its outbreak in 2003, was taken as an example for demonstration and evaluation in this study. The results showed that the new approach could help users to find and understand the contents and relationships among different information sources more efficiently. The keyword searching or natural language processing techniques could be used together with this approach to setup a better, well-integrated content database in the future.