REPOSITORY > RESULTS

Abstract

News spreads in many patterns, structures, and dynamics that change throughout time.
For a variety of reasons, certain news is only covered in a particular area. Language,
economy, geography, politics, time zone, and culture are just a few of the many barriers
that prevent news from reaching a larger audience. Observing these barriers reveals what
may influence the spreading of news reporting on different events.
The primary objective of this study is to develop methods and approaches to analyzing
news spreading barriers, with a particular emphasis on the above mentioned five barriers
(linguistic, economic, geographic, political, and cultural). The aim of the analysis of geographical
and time zone barriers is to identify the influence of time zone and geographic
position of news publishers on news spreading across the globe. Analysis of political barrier
is carried out to see the influence of political alignment of news publishers on news reporting.
The aim of the analysis of cultural and economic barriers is to identify the impact
of locations of news publishers with different cultures and economies on news spreading.
Analysis of the linguistic barriers is performed to determine the influence of publishing
language on news propagation.
This thesis focuses on three interconnected issues to the news spreading barriers. The
first issue involves adopting information cascade theory for news articles and event-centric
news dissemination analysis. The evaluation of the improved topic modeling and the
strategy for comprehending political and economic contrasts in news reporting constitute
the second issue. The third issue is to profile the news spreading barrier, a task to classify
news texts based on the stylistic choices of their news publishers.
A methodology is presented for the analysis of information propagation in news across
different barriers in different domains. To deal with the analysis of monolingual and multilingual
cascading analysis, different visualizations are incorporated. For all the five barriers
(linguistic, economic, geographic, political, and cultural), various analytical methods are
employed in this methodology. News related to three different domains (natural disasters,
climate change, and sports) is explored. The findings revealed that 1) the scope of a specific
event significantly affects the news spreading across languages, 2) the geographical
size of a news publisher’s country is directly proportional to the number of publishers and
articles reporting on the same information, 3) countries with shorter time-zone differences
and similar cultures tend to propagate news between each other, 4) news related to global
warming comes across economic barriers more smoothly than news related to FIFA World
Cup and earthquakes and 5) events which may in some way involve political benefits are
mostly published by those publishers which are not politically neutral.
A methodology is presented to comprehend the political, and economic disparities in
news reporting and to compare the sentiments expressed in various newspapers across
various geographic regions. In order to increase the quality of topics without changing the
fundamental structure of Latent Dirichlet Allocation (LDA), an improved topic modeling
technique is suggested that employs LDA with varying length of words and articles’ pooling
depending on queries. The COVID-19 news is used as an example in a variety of political
and economic situations. Our findings imply that news reporting by newspapers with
different political alignments supports the reported content. Also, economic issues reported
by newspapers depend on the economy of the place to which a newspaper belongs to.
An approach is presented for automatic barrier profiling using news meta-data. All
the data about news is obtained from the Event Registry global media monitoring system
and enhanced with metadata about the publishers fetched from Wikipedia. To annotate
the news articles across different barriers, annotation procedures are set. To deal with the
similarity between metadata of news articles, Euclidean distance is calculated and automatic
annotation is carried out. Barrier classification is performed using news headlines,
common sense inferences and sentiments. Evaluation is carried out using machine learning
classical classification methods, deep learning and transformer-based methods.