Discovering users navigation of online newspaper using Markov model
Husna Sarirah Husin, Norsuhaili Seid · 2017
Online newspaper readership is gradually increasing as users are actively spending more time on the Internet to read news. Understanding how users navigate the site is important so we can improve their Web experience. In this study, we collect Web server logs of an online Malaysian newspaper for 28 days in April 2012. We apply Markov model which is one the techniques in Web usage mining used to model a collection of user sessions. The content pages are categorized to three types; main page, article page and section page. We analyzed the navigation flow between different pages in a session. From the Markov state model, we discovered that users are inclined to continue reading the articles from the main page. Interestingly, majority of users that start navigation with the main page will also end their session with the main page as well. As for the section page, we investigate what section pages are read together in a session. Our findings also found that users tend are likely to read section pages together and we observed that National sections are always the favorite section, while Education is the least section to start their session. The findings from this study can be a basis for recommending pages to the user so they can navigate more pages in a session and in turn to increase traffic to the online newspaper site.