Time series analysis of some crime data in Nigeria

Abstract

This research compared and assessed various time series models applied to analyze some crime data in Nigeria. Four distinct time series trend models namely - linear, quadratic, exponential, and s-curve were fitted using annual data spanning from 1993 to 2022. The findings revealed that the quadratic trend model demonstrated the lowest MAPE and MAD, offering a suitable fit for murder, armed robbery, and assault crime data. In contrast, the exponential trend model proved most effective in accurately representing bribery and corruption crime data. It was suggested that the
quadratic trend model should be used to study the trend pattern of crime such as murder, armed robbery and assault in the long-term run, while the exponential trend model be used to predict the future behaviours of the bribery and corruption crime in Nigeria. Furthermore, the findings obtained from this study should be a catalyst for additional research and practical implementations within the realm of crime analysis and prevention.