A Crime Rate Forecast and Decomposition Method

Quanbao Jiang, Jesús Javier Sánchez Barricarte


Crime forecast is a hotspot in criminology. This paper comes up with a new stochastic crime rate forecast method, namely based on historical age-specific crime rates, we employ SVD (singular value decomposition) in matrix theory to lower the rank of the crime rate matrix, and then transform the time-series vector to a time-series variable problem, then we use time-series analysis to forecast the time-series variable and then the age-specific crime rates. With the forecasted age-specific crime rates and population projection, we obtain the forecasted crude crime rate and then decompose the difference between two crude crime rate into the change in age-specific crime rates, change in age structure and change in sex structure.

Full Text:


International Journal of Criminology and Sociological Theory | ISSN : 1916-2782