School Shooting : Threat Detection and Classification in Textual Leakage

• Main Author: Khan, Ajmal
• Format: Others
• Language: English
• Published: Uppsala universitet, Institutionen för informationsteknologi 2013
• Online Access: http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-207622

The continual occurrence of school shooting incidents underscores the need of taking preventive measures. Inductive measures of threat assessment have proved to be a bad strategy to solve the problem and new research is focusing on deductive approaches. Deductive threat assessment approaches are gaining ground and efforts are underway to mine text for automatic detection of threats in written text. Automatic detection and classification of threats in the digital world can help the decision makers focus energies on imminent threats of school shooting and take preventive measures in time to save precious lives and other resources. The contribution of this study is criticism of the previous work done on the problem of school shooting, collection of data of previous cases of school shootings in order to find out the factors that affect the school shooting problem and the development of an algorithm that could be used to detect threat of school shooting in written text in the English language. The algorithm proposed in this study classifies text on the basis of seriousness of the threat of school shooting in to four categories i.e., "High", "Medium", "Low", and "Not a threat". The seriousness of the threat is decided based on different indicators present in the text of the threat and presence of factors that has affected previous school shooters. A prototype is implemented to demonstrate the classification in to the categories mentioned above.