Algorithmic criminology

Abstract

Computational criminology has been seen primarily as computer-intensive simulations of criminal wrongdoing. But there is a growing menu of computer-intensive applications in criminology that one might call " computational, " which employ different methods and have different goals. This paper provides an introduction to computer-intensive, tree-based, machine learning as the method of choice, with the goal of forecasting criminal behavior. The approach is " black box, " for which no apologies are made. There are now in the criminology literature several such applications that have been favorably evaluated with proper hold-out samples. Peeks into the black box indicate that conventional, causal modeling in criminology is missing significant features of crime etiology.