PREDICTIVE BASED ANALYSIS OF CRIMINAL DATA (FIRS)

Abstract

Uses of Information Communication Technology (ICT) have significantly changed the working of almost every field. Law enforcement authorities also use these technological developments to maintain the public security. Data regarding criminal activities is collected and stored in large quantity which needs to be analyzed for finding different patterns. This analysis unlocked many secret of stored data which are many times ignored. Different Data Mining (DM) techniques have been developed for analyzing these data to find spot trends. So, there is a need to implement data mining techniques on criminal data to identify the crime patterns which are significantly helpful for law enforcement agencies to plan and control the security. In this paper, a complete analysis is performed by using predictive analytics techniques to analyze the criminal data to discover the crime trends. A framework has been purposed for predictive based analysis which is able to process large amount of criminal data and identify the hidden crime pattern and trends. Data is collected from local police to extract the important attributes. Purposed model used clustering and classification techniques, k-mean, J48 Decision Tree and Naive Bayes (NB) to explore the hidden crime profiling. These explorations are used for future prediction of crime occurrences.  Later these predictions will facilitate the law enforcement authorities while making the security policies and deployment of security personals.

Keywords-component; Data mining; Classification; Clustering; Crime; Security; Naive Bayes