A PREDICTIVE MODEL FOR MAPPING CRIME USING BIG DATA ANALYTICS

Abstract

Crime reduction and prevention challenges in today's world are becoming increasingly complex and are in need of a new technique that can handle the vast amount of information that is being generated. Traditional police capabilities mostly fall short in depicting the original division of criminal activities, thus contribute less in the suitable allocation of police services. In this paper methods are described for crime event forecasting, using Hadoop, by studying the geographical areas which are at greater risk and outside the traditional policing limits. The developed method makes the use of a geographical crime mapping algorithm to identify areas that have relatively high cases of crime. The term used for such places is hot spots. The identified hotspot clusters give valuable data that can be used to train the artificial neural network which further can model the trends of crime. The artificial neural network specification and estimation approach is enhanced by processing capability of Hadoop platform.