Charla de Tesis II Magíster en Ciencias mención Computación/ Alumno: Hernán Sarmiento
Título Tesis: "Detecting Emergency Situations Based on Locations in Twitter".
Resumen: Social media has played a critical role in the last fifteen years to detect, summarize and classify messages during emergency situations.
Most of the methods that use Twitter for detecting emergency situations rely on keywords. The problem with keyword-based methods is the need to train models in specific domains, in multiple languages, and for different types of events, for example: earthquakes, terrorist attacks, etc. In contrast, our proposal uses the recurring references of a location in the metadata of microblogging messages in order to identify such events without using keywords and characterize them through probability distributions of the interarrival time of the messages.