| 摘要: | 本論文中，提出不同於內容過濾與協同過濾之推薦系統，以社會導覽(Social Navigation)為基本概念，將群組內使用者瀏覽過的網頁進行推薦。對系統而言，利用初步的過濾以降低後端儲存以及計算的時間。
在使用者訓練中，首先透過訓練，分析使用者瀏覽行為。利用訓練後使用者所獲得本身興趣的所屬類別，作為網頁投票的依據。在網頁的過濾上，透過使用者的瀏覽行為，將所蒐集的網頁進行所屬類別的投票，進而推薦。推薦上，以toolbar 作為與使用者的溝通方式。讓使用者在瀏覽器上安裝toolbar，就可以得到其他使用者的推薦資訊。
本論文的貢獻在於能改善Collaborative Filtering在處理效率上缺點，提供使用者對於自己感興趣的網頁類別中，其他使用者具有相同類別的網頁資訊。
In the following research, using the concept of social navigation, the web pages browsed by users are used to recommend web pages to each other.
After the analysis of the history of recently viewed web pages, we can learn more about the user’s interests. In the filtering of web pages, we can set up a voting method in the target class mechanism, to recommend web pages to other users that have similar interests. The toolbar is the main medium of communication with the users. The user can immediately get recommended web pages on the browser.
The main contribution of this paper is to improve collaborative filtering, and to provide web pages for other users with similar interests. As for the system, the initial filtering would lower the time required for calculation as well as the amount that needs to be saved. |