Search
Program Calendar
Browse By Day
Browse By Person
Browse By Room
Browse By Category
Browse By Session Type
Browse By Research Area
Search Tips
ASC Home
Personal Schedule
Sign In
Session Submission Type: Complete Thematic Panel
Machine learning techniques offer the promise of advancing the field of crime at place by providing superior predictive techniques. However, the results often are a “black box” that do not yield insights of interest to criminologists interested in the etiology of crime. The papers on this panel explore opening this black box in various ways in various settings using various machine learning techniques. The results indicate a useful interchange between these black boxes and possible theoretical development.
What Do Deep Learning Models Show Us About Crime Attractors - George Mohler, Indiana University-Purdue University Indianapolis; George Tita, University of California, Irvine; Chris Goebel, Indiana University-Purdue University Indianapolis
Opening the Black Box of Machine Learning: Demonstrating Nonlinear Interactions Using KRLS in Studying the Spatial Location of Crime Events - Madison Huang, University of California, Irvine; Christopher J. Bates, University of California, Irvine; John Hipp, University of California, Irvine
Tracking Changes in Crime-Feature Correlation Through Bayesian Data Assimilation - George Mohler, Indiana University-Purdue University Indianapolis; Martin Short, Georgia Tech University
Forecasting Police Calls for Service with Graph-Based Deep Neural Networks - Jeffrey Brantingham, University of California, Los Angeles; Bao Wang, University of California, Los Angeles; Xiyang Luo, University of California, Los Angeles; Baichuan Yuan, University of California, Los Angeles; Fangbo Zhang, University of California, Los Angeles; Andrea L. Bertozzi, University of California, Los Angeles