Exploring the fractal nature in dynamics of crimes during recent Lok Sabha elections in West Bengal

Abstract

This work aims towards studying the dynamics of crime rate at election times. It is quite expected that the crime rate at normal times would fluctuate exhibiting some dynamics that may change drastically at times of elections. The consolidated report on the number of arrests made has been aggregated from the public domain on the strength of an ongoing research. The temporal behaviour of the dataset so obtained is then subjected to frequency domain analysis. Information entropy is considered as a metric for studying the uncertainty prevailing in terms of arrests. Fractal dimension changes in a moving window has been performed to assess the self similar nature of the dynamics. This brings out quite interesting patterns at around the times of general election. This carries some lessons for controlling law and order situation during turbulent election times.