以可延伸標記語言建立一個自動化文件管理系統

Abstract

當文件數量大時，如何管理這些文件是很重要的。在本篇論文中，我們提出一個三層式的自動化文件管理架構，包含呈現層、邏輯層、及資料層。為了容易提供個人化服務及讓資料顯示在多種裝置上，在本系統中，所有的資訊皆以可延伸標記語言（XML）的格式儲存。並且，透過我們可延伸標記語言的設計，可以達到以下三點特性：容易存取目錄結構、容易新增/刪除文件、容易儲存細分/合併目錄的結果。另外，我們亦提出自動分類演算法，因此，當新增一文件時，系統會自動將此文件分配到最適當的目錄中;而當某一目錄中的文件數目太多時，系統會自動在此一目錄下建立新子目錄，並將目錄中所有的文件重新分配到最相似的目錄。為了在相關性高的目錄間自動建立相關性鏈結，我們亦提出一個自動化建立相關性鏈結演算法。最後，我們依據設計好的系統及演算法實作一個以可延伸標記語言為基礎的自動化文件管理雛型系統,展示了實際應用上的可行性。結果顯示利用本雛型系統，可達到自動化管理文件及目錄、容易提供個人化服務、及容易將資料顯示到不同的裝置上。

As the amount of documents becomes large, how to manage these documents becomes a very important issue. In this thesis, a three-layer Smart Document Management system, which is composed of Presentation layer, Logic layer, and Data layer, is designed. In order to provide personal services and display data to multi-device easily, all information of documents and directories is stored in XML format in our system. In addition, by our XML design, it can be very convenient to access the structure of directories, add/delete documents, and store results of splitting/merging directories. We also propose clustering algorithms, so that when a new document is inputted, it can be automatically assigned to its most similar directory, and when a directory contains too many documents, it can automatically generate new subdirectories and reassign each document of the directory to its most similar directory. Furthermore, in order to automatically establish associative links between related directories, we propose Directory Associating algorithm. Finally, a prototype of the Smart Document Management system has been implemented to show the system and the algorithms proposed in the thesis are workable. The results show that documents and directories can be managed automatically, and that providing personalized services and displaying identical data into multi-device are very easy and convenient.