Context Oriented Analysis of Interest Reflection of Tweeted Webpages based on Browsing Behavior

Abstract

Twitter, the most popular micro-blog, attracts more and more Web users to share their accessed webpages and stimulates diverse recommendation mechanisms on social Web. However, the difference between webpage access and sharing on Twitter is often ignored, and many recommendation mechanisms are proposed based on an unproven common sense: "share" well reflects "interest" (users share their favorite webpages containing interested contents after accessing them). In this paper, we explain the difference between webpages access and sharing by giving possible reasons, and confirm them with actual users' activity data. We study the browsing behavior and develop a novel context-oriented approach to deeply analyze interest reflection of tweeted webpages by integrated using net view data, twitter data, and webpages. The experimental result shows our approach can effectively evaluate credibility of interest reflection on Twitter.