透過您的圖書館登入
IP:3.137.172.68
  • 學位論文

以術語分類從網路客戶端探求個人化偏好

A Term Categorization Approach to Discovery of Personal Preferences from Web Clients

指導教授 : 簡立峰

摘要


個人化的應用在網路世代一直是一個相當重要的研究方向,透過個人化的服務能更貼近使用者的需求,本研究旨在運用術語分類的技術從網路客戶端的資料來源中探求使用者在獲取資訊時所擁有的偏好或興趣,要決定使用者有哪些特殊偏好,我們首先從客戶端(例如個人電腦)中的大量儲存的歷史文件資料作抽詞的動作,本研究著重於三項資料來源:使用者搜尋關鍵詞、瀏覽過的網頁及電子郵件,之後我們將整理出來的關鍵術語再透過術語分類的技術來決定術語的分類目錄,以建立使用者各個不同百分比的目錄分布,本研究對五個志願使用者作實驗,從各個不同資料來源來產生目錄分布,並且對產生出的目錄分布做簡單的觀察分析。 這項簡單的方法並未涉及太深的分析與架構,在各階段的轉換還有很多的討論空間,例如個人目錄的分布轉換到個人偏好的探討,但我們相信這是一個有潛力的方法去實踐各種不同的個人化應用,例如個人化搜尋、個人資訊過濾與推薦系統。

並列摘要


This thesis is developed for discovering the personal preferences from different data sources that stored in web client side like personal computers. To determine the personal preferences, we first extract key terms from different data sources to be the material of the next step, there are three major data sources:search keywords, browsed pages and e-mails. And then we use term categorization method on the key terms to generate user profiles that contains the distributions of different categories. After the simple method, we do some observations of five users to discuss the appearance of the personal categories. This method is really simple and has more space to discuss about the transformation between term categories and personal preferences. We hope the simple method can be a potential way to perform personalization applications, like personalized search, personal information filtering or recommendation system.

參考文獻


[1] Staurt E. Middleton and David C. de Roure. Ontological User Profiling in Recommender Systems. In ACM Tansaction on Information Systems 2004
[2] Kazunari Sugiyama. Adaptive Web Search Based on User Profiles Constructed without Any Effort from users. In www2004
[11] Gediminas Adomavicius. User Profiling in Personalization Applications through Rule Discovery and Validation. In KDD 99
[13] J Teevan, ST Dumais, E Horvitz. Personalizing search via automated analysis of interests and activities. In Proceedings of SIGIR, 2005.
[14] Google suggest. http://www.google.com/webhp?hl=en&complete=1

延伸閱讀