Ontology-Based Terrorism Event Extraction

The proliferations of terrorism news articles from thousands of different sources are now available on the Web. Summarization of such information is becoming increasingly important. The aim of this paper is to study and compare the linguistic feature methods that are appropriate for use in terrorism event extraction systems. The event extraction has a main function to named entity recognition and segments the terrorism events from news articles to display to the users. The research methodology in the paper compares many linguistic features techniques including the terrorism gazetteer, the terrorism ontology and terrorism grammar rule. The annotated entities are summarized into the three desired template events. The terrorism events are classified by using similarity measure based on Term Frequency-Inverse Document Frequency called TF-IDF-based event segmentation. Additionally, we use a finite state algorithm to learn these feature weights and also studied to emphasize the performance of the event extraction algorithms. The experimental results show that the terrorism ontology linguistic feature selection yielded the best performance with 85.15% for both precision and recall.