個人知識重整-以記憶模型與標籤技術為基礎

Abstract

資料夾為自從電腦開始使用以來，就一直擔任著整理資訊的工具，但是資料夾本身有其限制及缺點，例如：樹狀結構所產生的問題、階層優先度問題等，導致有時候無法有效地尋找到所需的資訊，造成人們的困擾。現有的檔案搜尋系統，如Google Desktop Search，限於關鍵字與資訊內容可能不太相關的問題，也無法完全解決這個問題。由於標籤網路為平面式網路結構無階層性問題，以及是由使用者對資訊的認知所下的標籤，所以可以避免檔案搜尋系統關鍵字相關度的問題，所以我們將採取此結構來做為解決方案。 本論文以在網際網路上發展迅速的標籤(Tag)分類法，針對個人組織整理其所擁有資訊的方面，來設計出一個個人知識重整系統，以認知心理學的記憶模型和記憶歷程的理論為基礎，期望能改善資料夾和檔案搜尋系統現有的問題，以及進而能幫助人類做記憶中知識的重整。針對高中三年級的學生，藉由採用個人知識重整系統和傳統資料夾分類二種方式來做實驗，第一階段請二班各別使用二種方式來整理圖檔，隔週之後，第二階段請學生使用當初建立結構的系統來找尋指定的目標圖片，於建立與搜尋的過程中以電腦程式記錄下使用者的動作。比較二者搜尋到目標檔案的效率、各自結構建立的風格、二種結構的搜尋風格差異，以及個人與整體間是否存在某些認知的共識。 研究結果發現使用個人知識重整系統比資料夾在搜尋目標檔案時，時間上較為快速，步驟數較多。使用者大部分為圖片加上2~3個標籤。個人所建立的標籤網路，在某些部分會和團體共同認可的結構相同。

Since the first version of Window operation system was developed in 19XX, a categorization approach “folder” have served as an important information tool to organize data. However, the hierarchical structure of folder has its limitations and weaknesses for organization data, e.g. problem of tree structure and hierarchical priority…etc. The drawback of folder categorization may make people difficult to effectively search for demanded information. Besides, the existing file searching system like Google Desktop Search limited by keyword may not reflect the context of content, so it needs a more effective and efficient approaches to help people search in organized knowledge structure even unorganized data pool. In this study, I focus on re-organizing knowledge by tagging approach to help data search and organization. Because tagging network is structured as plane network, it has no hierarchical problem. And due to tagging procedure is established based on user’s cognition, users can avoid content relation problem resulted from keyword search. This thesis uses tagging technology to design a personal knowledge re-organizing system which is based on memory models and memory recording processes of cognitive psychology. The proposed tagging system aims to improve original files searching system based on folder structure, and thus can help people record, store, and re-organize knowledge in user’s personal biological memory. Several experiments were run in senior high school for evaluating the proposed tagging system by comparing the traditional hierarchical folder system. Two phases experiments including pretest and posttest show that using tagging system to re-organize personal knowledge is more efficient, effective, and intuitive than traditional approach. Moreover, by visualizing how user records and organizes his/her knowledge, the tagging network structure organized by different user can show user’s personal memory style which can be viewed as a useful tools to observe how people organize their personal knowledge.