TY - GEN
T1 - On the Analysis of Information Resilience and the Spread of News
AU - De Oliveira Capela, Fernanda
AU - Martinez-Mejorado, Denisse
AU - Ramirez-Marquez, Jose
N1 - Publisher Copyright:
© 2022 IEEE.
PY - 2022
Y1 - 2022
N2 - This paper presents TopicRes, a five-step methodology to analyze the impact of real-life events as a function of how topics spread in online news media. Combining concepts such as text analytics, network modeling, and systems resilience, TopicRes assists in exploring and analyzing online news media spread. Data analytics and statistical tools are used in every step: data collection, topic detection, topic influence network, and topic resilience. We present a case study in the Portuguese language to showcase how TopicRes functions, its usefulness, and its versatility. The results show that the system resilience model is convenient for identifying events and capturing their dynamic behavior over time. The network deepens this analysis with detailed static snapshots of the topics and their relationships. Results are then clustered by behavior, presenting a new way to fathom the system's dynamics enduring a certain type of disruptive event. In the case study, it is possible to observe the power dynamics of the media outlets and how the local structure influences the news spread. TopicRes is a powerful analytic tool to sense important events in the media, aid in disaster response and crisis management, track the development of new technologies, and 'fake news' propagation.
AB - This paper presents TopicRes, a five-step methodology to analyze the impact of real-life events as a function of how topics spread in online news media. Combining concepts such as text analytics, network modeling, and systems resilience, TopicRes assists in exploring and analyzing online news media spread. Data analytics and statistical tools are used in every step: data collection, topic detection, topic influence network, and topic resilience. We present a case study in the Portuguese language to showcase how TopicRes functions, its usefulness, and its versatility. The results show that the system resilience model is convenient for identifying events and capturing their dynamic behavior over time. The network deepens this analysis with detailed static snapshots of the topics and their relationships. Results are then clustered by behavior, presenting a new way to fathom the system's dynamics enduring a certain type of disruptive event. In the case study, it is possible to observe the power dynamics of the media outlets and how the local structure influences the news spread. TopicRes is a powerful analytic tool to sense important events in the media, aid in disaster response and crisis management, track the development of new technologies, and 'fake news' propagation.
KW - Networks
KW - News spread
KW - Resilience engineering
KW - Text mining
KW - Topic modeling
KW - Topic resilience
UR - http://www.scopus.com/inward/record.url?scp=85151688323&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85151688323&partnerID=8YFLogxK
U2 - 10.1109/ICSRS56243.2022.10067823
DO - 10.1109/ICSRS56243.2022.10067823
M3 - Conference contribution
AN - SCOPUS:85151688323
T3 - 2022 6th International Conference on System Reliability and Safety, ICSRS 2022
SP - 343
EP - 352
BT - 2022 6th International Conference on System Reliability and Safety, ICSRS 2022
T2 - 6th International Conference on System Reliability and Safety, ICSRS 2022
Y2 - 23 November 2022 through 25 November 2022
ER -