Social media sites are playing a significant role in rapid propagation of information
when disasters occur. Because of the information that can be collected from these
platforms, Online Social Networks (OSNs) have gained a significant lot of interest in
recent years for their potential application in the spatial and temporal modeling of
events. One of the most untapped uses in this area is natural disasters monitoring.
OSN users' vital information can help with relief operations both during and after a
disaster. Social media, especially Twitter, is increasingly being used to enhance
response to extreme weather events / emergency management scenarios, such as
floods, by publicizing potential risks and their consequences, as well as educating
authorities. In this thesis a system has been developed that can identify tweets related
to natural disasters. In order to be able to implement this system, various machine
learning algorithms were used, such as the Naïve Bayer or the decision trees
algorithm and depending on the results, the most appropriate one was selected.
Logistic Regression (LR) and Support Vector Machines (SVM) algorithms had the
best results. The contribution of this dissertation is the development of a new method
for services and researchers dealing with natural disasters to swiftly recognize an
emergency scenario using data from Twitter.
Collections
Show Collections
