Detection of Bias in Media Outlets with Statistical Learning Methods - Text Mining: Classification, Clustering, and Applications - page 45

Jazeera in the way they report the same events. Although this finding is far
from surprising, the fact that it could be done in an automatic way by using
statistical learning algorithms has many implications: a large scale implemen-
tation of this system could easily be used to monitor a large number of news
outlets, and perhaps cluster them according to their similarity in topic / term
biases. This in turn could help us to identify different accounts of the same
story, that is accounts coming from news outlets that have a significantly dif-
ferent bias. Having access to different versions of the same story is of course
a valuable opportunity, as it can help us to form an opinion as independent
as possible about current events. Then we have also presented a method
to compute distances between media outlets, based on their term-choice and
topic-choice biases.
Recent related work, discussed above, points in the same direction. The
detection of different political perspectives in authors, groups, or speakers
has attracted significant attention of the NLP community, and is also partly
related to the task of opinion analysis. Its large scale application to media
analysis can truly change that field of scholarship.
Despite the small scale of this case study, we feel that modern AI technology
has an important role to play in media content analysis, as well as in the social
sciences. When scaled up to include hundreds or thousands of media outlets,
a goal easily achievable also with standard equipment, these methods can lead
to informative maps showing the relation between media outlets based on the
analysis of statistical patterns in their content.
References
[1] Aljazeera News, http://english.aljazeera.net/
[2] Brank, J., Grobelnik, M., Milic-Frayling, N., and Mladenic, D. Feature
selection using support vector machines . Proc. of the Third International
Conference on Data Mining Methods and Databases for Engineering,
Finance, and Other Fields, 2002.
[3] CNN News, http://www.cnn.com
[4] Cristianini, N. and Shawe-Taylor, J. An Introduction to Support Vector
Machines and Other Kernel-Based Learning Methods . Cambridge Uni-
versity Press, 2000.
[5] Detroit News, http://www.detroitnews.com
[6] Fortuna, B., Grobelnik, M., and Mladeni, D. Visualization of Text Doc-
ument Corpus . Informatica 29 (2005), 497-502.