| Author: | Nick Shapiro |
| Advisor: | Charles Owen |
| Email: | shapiron@cse.msu.edu |
Website navigation can be a tricky thing to set up. A webmaster designing a site may think that visitors would likely be interested in certain pages in a certain order, but the interests of the actual visitors may be totally different. There has been a recent trend on the Internet of "popularity engines" that allow users to follow in the footsteps of the most popular choices by previous users. This study is an attempt to apply the popularity concept to website navigation. The weblogs of a large, very popular website are examined and Markov Models are applied in an attempt to predict what pages visitors will be next interested in, based on the habits shown in previous visits. If prediction proves to be effective, links can then be "juggled" and "advertised" to ease navigation for the user.
Return to Workshop Listing