Explore a research presentation examining how differences in crime reporting by victims impact predictive policing systems. Delve into the study conducted by N. Akpinar, M. de-Arteaga, and A. Chouldechova, presented at the FAccT 2021 conference. Learn about the background of predictive policing, previous work in the field, and the model developed by the researchers. Understand the use of synthetic data and hotspot prediction techniques. Analyze the results, including rescaling methods and the moving average model. Gain insights into the conclusions drawn from this important research on the intersection of crime reporting, predictive algorithms, and potential biases in policing systems.
The Effect of Differential Victim Crime Reporting on Predictive Policing Systems
Association for Computing Machinery (ACM) via YouTube
Overview
Syllabus
Introduction
Background
Previous Work
Model
Synthetic Data
Hotspot Prediction
Results
Rescaling
Moving Average Model
Conclusions
Taught by
ACM FAccT Conference