以熱門主題趨勢分析為基礎之個人化部落格文章推薦

Abstract

由於資訊科技技術的快速進步，Web2.0已成為相當熱門的社群媒體發佈平台。於Web2.0所提供的各式應用服務中，部落格(blogosphere)提供相當便利的環境供使用者分享個人偏好與表達情感。除此之外，大部分的使用者也傾向於接收最新的訊息與熱門主題資訊，但隨著部落格文章與使用者的快速增加，使用者很難從大量的文章主題中發掘有興趣，或是熱門的文章內容。因此，我們認為部落格推薦系統中應考量文章主題的熱門程度。 本研究藉由預測主題之熱門程度，提出以熱門主題趨勢分析為基礎之個人化推薦系統。第一，我們分析部落格文章以辨識熱門主題，並使用部落格與Google搜尋透視所提供之資訊進行主題趨勢分析。第二，我們藉由分析使用者有興趣的文章內容(即使用者曾閱讀過的文章)、文章內主題之熱門程度與使用者對於相似文章的喜好程度，以瞭解使用者個人對於特定主題文章的偏好程度。第三，我們結合傳統推薦方法(即內容導向式過濾和以項目為基礎之協同過濾)與以熱門主題為基礎之推薦方法，提升推薦系統的效能與準確率。最後，實驗結果證明本研究提出之方法能夠依據主題熱門度與使用者興趣，有效率的推薦使用者喜好的文章。

Web 2.0 has become a popular social media on the Internet due to the fast evolution of the Internet technologies, resources, and users. Among the applications of Web 2.0, blogosphere is a new Internet social media for users to express their preference and personal feelings. Most of the people tend to receive the newest information and articles related to popular issues. However, with the rapidly increasing number of active writers and viewers, it is hard for people to discover useful information that is beneficial or interesting to them. Accordingly, it is necessary to develop a recommendation approach that takes the emerging popular blog topics into consideration. In this work, we propose a novel hybrid popular topic-based recommendation approach based on the predicted popularity degrees of topics to recommend blog articles to users. First, we analyze blog articles to identify popular topics, and then derive the popularity degrees of topics based on the blog-based popularity trend analysis and Google Insights-based popularity trend analysis. Second, our approach derive users’ personalized preferences on target articles of popular topics by considering user interests (article-push records), the predicted popularity degree of the topics, and user interests on the neighbors of target articles. Third, several recommendation methods are developed to enhance recommendation accuracy by combining our popular topic-based recommendation approach with the traditional filtering approach, i.e. content-based filtering (CBF), and item-based collaborative filtering (ICF). Finally, the experiment result demonstrates that the proposed approach can effectively recommend users’ desired blog articles with respect to topic popularity and personal interests.