Linked data for flood disaster

• Main Authors: Cortes, Hector Brian B., Manila, Andre Carlo F., Valdez, Roi Luigi
• Format: text
• Language: English
• Published: Animo Repository 2012
• Subjects: Flood control > Philippines; Flood damage prevention; Disaster relief; Computer Sciences
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/11050
• Institution: De La Salle University

In todays world, disasters have been increasing and improvements in the field of disaster management are needed to cope up with these changes. Many typhoons occur within the Philippines annually, thus its susceptible to major floods. Information about these typhoons are documented by news outlets and recorded in unstructured textual formats. To alleviate this kind of problem, this study proposes the utilization of information extraction to retrieve useful data contained within news articles and the creation of linked data to organize them into structures data that can be shared with other web resource. Finally, this program creates links with GeoNames and web portals of news outlets that would allow user to gain more information from those web portals. The proponents performed two testing one for the article classifier and another for information extraction. For article classification, the system was able to achieve an F-measure score of 80% and Accuracy of 83.33%. For information extraction, the system was able to achieve an F-measure score of 100% for typhoon name, 83.20% for affected area and 83.33% for typhoon properties.