Shared Conceptualisations in Weblogs paper
The final version of our BlogTalk 2.0 paper is available:
This is the abstract: In this paper we investigate how conceptualisations can be identified in weblogs using language technology (automated text analysis). We focus on getting a handle on both the concepts bloggers use and the way they think these concepts are related. The analysis of these conceptualisations can then be applied to a single weblog, resulting in a visualisation of potential conceptualisations the blogger wants to share with the outside world. Another type of analysis is to determine the overlap, or sharedness, of conceptualisations between bloggers. We have implemented both analysis approaches in an interactive tool [called Sigmund].
Work on making Sigmund publicly available will now resume. The main bottleneck is ensuring that the weblog spider is reliable enough. The relevant paragraph of the paper is:
The largest practical problem we encountered is the availability of a weblog spider as outlined in Section 3.1. Our current version of such a spider is reasonably reliable, but by no means perfect. The only persistent solution appears to be that blogging software provides a public interface with access to all posts of a blog in a preferably established format such as RSS. Obviously, weblog research in general would greatly benefit from this.
Comments