Archives

Crime Data Clustering Using Neighbourhood Rough Set


Lydia J Gnanasigamani and Seetha Hari
Abstract

Analysis of crime data has been historically done to find the patterns and associations among criminal incidents. Many faces of crime such as violent crimes, child-related crimes, property offences and the reasons for such crimes like the socio-economic status, education level, the immigration population level in the society have been researched extensively. Similar crime can be classified under multiple categories according to the practices of the police officers of that area. Clustering is an unsupervised classification method which uses the characteristics of the data to group it into clusters. So clustering crime based on its attributes will help in similar crimes be grouped together and enable the identification of hidden patterns. In this work, we cluster the crime data based on neighbourhood roughness of the attributes. This ensures that closely related crimes fall under the same cluster. Moreover, the proposed algorithm works for both numerical and categorical attributes without the requirement of data conversion or discretization.

Volume 11 | 11-Special Issue

Pages: 478-484

DOI: 10.5373/JARDCS/V11SP11/20193056