In order to join virtual sessions, you must be registered and logged-in(Were you registered for the in-person meeting in Denver? if yes, just log in.) 
Note: All session times are in Mountain Daylight Time.

Learning crime patterns from victimisation-embedded network

Authors: Tongxin Chen*, SpacetimeLab, University College London, Yang Zhang, SpacetimeLab, University College London
Topics: Urban Geography
Keywords: Crime patterns, network analysis, graph, unsupervised learning.
Session Type: Paper
Presentation File: No File Uploaded

What is significant for place-based crime prevention is putting the protective practices to the victim who has a vulnerability and the place with high risk together to reduce the individual vulnerability of victimization. Though network analysis has been widely applied in detecting the offender network (e.g., co-offending network, organized criminal network), a victim-based network analysis seldom attracts attention in such empirical studies.

This paper explores the theft patterns approached by a victim-based network analysis with an interesting question: How could the victimization-embedded network explain the theft pattern in an urban area. We propose a victimization-embedded network referring to the victims connected by their similarity of the victimization represented by the victim’s demographics (i.e., gender and age level) and target type. Then we utilized a community detection algorithm to several significant “victimization communities” in the network. Based on the detected community on the network, we analysis their difference in spatial and temporal patterns, as well as the connection among communities and their physical environment indicators.

Our experiments on the police record data from Hangzhou, China show that distinct differences in spatio-temporal patterns of theft offense among the communities detected from our network, which could inspire the crime prevention strategies can be integrated to some certain types in some places based on the detected victimization community.

Abstract Information

This abstract is already part of a session. View the session here.

To access contact information login