Easy access to the Internet, and to free blog hosting services, has led to growth in the creation of blogs. Millions of blogs are available on various topics, ranging from travel, parenting, and cooking, to healthcare and investment. Just about anyone can start a blog these days and write about a particular topic of interest. But what do we know about the reliability of the information provided in blogs? What is the credibility of the blog writer or of the blog itself?

This paper discusses the use of a credibility framework to condition the retrieval of blog posts. The proposed framework is based on Rubin and Liddy’s work [1]. Credibility indicators are divided into post-level and blog-level indicators. Post-level indicators include parameters such as spelling mistakes, length of the post, timeliness, and text quality. Blog-level indicators include parameters such as expertise of the blog writer, regularity, topical consistency, regularity of the blog, and number of comments received on blog posts. The authors propose two approaches for incorporating credibility indicators into blog post retrieval: credibility-inspired re-ranking and combined re-ranking.

The paper provides detailed explanations of these credibility indicators, including mathematical formulas that demonstrate their use. The experimental setup is well explained and the obtained results are well tabulated, accompanied by in-depth analysis of the example blogs and posts. On the whole, the paper is exhaustive, well organized, and easy to understand. It makes an interesting and constructive read for those who are pursuing research in the area of credibility in blogs. I would also recommend this paper to people who write blogs, to get an idea of the features that make a blog a credible.