ICTlogy

Abstract:

Just as there has been a growing interest on personal learning environments as useful learning tools, interest should be also paid to “personal teaching environments”, light open source content management systems (CMS), such as WordPress or Drupal, that, configured in an appropriate way, allow a teacher to set up quickly and easily an open environment for a virtual classroom. In this paper we describe a reputation scheme that can be used to rank resources in a blog (posts, comments and even other users) according to the interaction they generate. Most blogs allow users the possibility of browsing recent posts, but finding older posts (and other blog contents) can be a difficult task. Using the proposed reputation scheme, users would be able to find the most interesting resources, thus breaking the imposed timeline of a blogbased light CMS. Our experiments show that the number of comments and votes is not enough for determining the importance of a post (or a comment). On the other hand, the structure of the graph (i.e. the number and depth of the branches representing discussions around a post) is a more interesting indicator of post relevance. In order to do so, we have manually labeled each post according to its importance and we have tested several measures trying to determine the “optimal” one, which is based on graph path lengths. Then we have tried to compute a weighted metric combining some of these measures using principal component analysis. Our results show that we can automatically detect the most relevant posts and comments. Once implemented as a WordPress plugin, we expect this reputation scheme to help learners to better follow and engage into blogbased open courses.