Detection of Bias in Media Outlets with Statistical Learning Methods - Text Mining: Classification, Clustering, and Applications - page 43

the mapping based on this distance can be seen in Figure 2.4.
FIGURE 2.4 : This plot shows the relative distance between news outlets,
using the topic similarity described in the text.
Again, here Al Jazeera is distant from the others. A possible interpretation
could be that this reflects the effect of regional focus.
2.7 Related Work
In recent years there has been significant attention to various problems of
text analysis that are related to the ones discussed here. Three papers ap-
pearing in 2006 (roughly at the same time in which the experiments reported
here were under way) are worth mentioning, as they have both significant
connections and interesting differences.
The articles (10) and (11), partly involving the same authors, focus on the
concept of 'perspective' from which an article is written. In the first article,
the authors use a dataset created using the “bitterlemons” website: a debate
website set up to contribute to mutual understanding between Palestinians
and Israelis by publishing two views of each topic.
In the second article, two of the same authors analyze the same 'bitter-
lemons' dataset as well as another dataset, formed by “transcripts of the
three Bush-Kerry presidential debates in 2004, provided by the Commission
on Presidential Debates. Each document is roughly an answer to a question
or a rebuttal.”
One elegant feature of these studies is that by construction the two datasets