General Crime Mining Framework for Unstructured Crime Data Prediction

The high volume of unstructured data of crime reports in police stations and the complexity of relationships between these data has made the importance of crime analysis is increasing day after day. This paper proposed a general crime mining framework that integrates the text mining techniques with the crime analysis process to improve the policemen decisions and reduce their efforts. The proposed framework is used in this paper to predict the crime type of the vast amount of unstructured police incident reports, which considered an important part of the criminal behavior analysis. The primary evaluation of the result can predict the crime type with accuracy 85%.