Crime Type Document Classification from Arabic Corpus

This paper describes an initial prototype for identifying types of crime in a text within the crime domain. Two approaches are explored to perform recognition tasks. The first approach completely relies on direct recognition using gazetteers. In this case, lists of crime verbs and crime names are used. The second approach is a rule-based system. Rules are built based on the predefined crime indicator list that contains some important keywords. Even though the system is still under development, the initial results are promising.