以熱門事件趨勢分析為基礎之部落格文章推薦

Abstract

部落格(Blog)是許多Web2.0應用中相當熱門的網路社群，許多使用者可以撰寫文章分享資訊或是瀏覽文章。部落格文章通常包含有事件相關的資訊，透過分析落部落格文章可分析出事件的相關趨勢，而使用者通常會透過部落格來閱讀有興趣的熱門事件文章。隨著網路快速的發展，過多的部落格文章造成資訊過載的問題。因此，本研究提出以熱門事件趨勢分析為基礎之部落格文章推薦方法，包括提出結合部落格與Google Insights的事件趨勢分析方法，以找出具有熱門趨勢的事件，並提出整合熱門事件、文章熱門程度與內容式過濾之個人化部落格文章推薦方法。實驗結果顯示本研究所提出的方法比傳統方法能更有效的針對使用者的興趣來推薦適合的部落格文章。

Web 2.0 is a new way of the Internet, and brings many applications. Among Web 2.0 applications, weblog have emerged as a new communication and publication medium on the Internet for diffusing the latest useful information. Blog articles usually contain event-sensitive information, so they represent the opinions of the populace and constitute a reaction to current events on the Internet. Although blog articles contain much information about event, with the huge growth of bloggers and blog articles, blog readers are difficult to find interested articles. Accordingly, it is important to provide recommendation service for people in the blog platform, especially for recommending blog articles of emerging or popular events that suit their interests. In this work, we propose novel personalized event-based blog article recommendation approaches, which combines the popular event-based trend analysis and personal preference to recommend blog articles of popular events that suit user interests. The event-trend analysis analyzes the popularity trend of events, and predicts the popularity of events based on the blog-based popularity trend analysis and Google-based popularity trend analysis. Moreover, we further propose a novel approach to infer users’ preferred popular events based on the predicted popularity degree of events and users’ personal interests. The event-based approach recommends blog articles by integrating personalized popularity of events, content-based filtering and event-based popularity of blog articles. Our experimental results show that the proposed approaches outperform conventional approaches.