Forecasting Crime Using ARIMA Model

TL;DR

This paper demonstrates how ARIMA time series models can effectively forecast future crime rates in London boroughs using historical crime data, aiding police decision-making and crime prevention strategies.

Contribution

It introduces a methodology for crime rate forecasting using ARIMA, showing improved accuracy over exponential smoothing with real London crime data.

Findings

ARIMA achieved 80% accuracy in crime prediction.

ARIMA outperformed exponential smoothing in model fitting.

Forecasting with 5 years of data predicted 2 years ahead.

Abstract

Data mining is the process in which we extract the different patterns and useful Information from large dataset. According to London police, crimes are immediately increases from beginning of 2017 in different borough of London. No useful information is available for prevent crime on future basis. We forecasts crime rates in London borough by extracting large dataset of crime in London and predicted number of crimes in future. We used time series ARIMA model for forecasting crimes in London. By giving 5 years of data to ARIMA model forecasting 2 years crime data. Comparatively, with exponential smoothing ARIMA model has higher fitting values. A real dataset of crimes reported by London police collected from its website and other resources. Our main concept is divided into four parts. Data extraction (DE), data processing (DP) of unstructured data, visualizing model in IBM SPSS. DE extracts crime data from web sources during 2012 for the 2016 year. DP integrates and reduces data and give them predefined attributes. Crime prediction is analyzed by applying some calculation, calculated their moving average, difference, and auto-regression. Forecasted Model gives 80% correct values, which is formed to be an accurate model. This work helps for London police in decision-making against crime.