Mining for Spatial Associations in Urban Residential Auto Theft Analysis

Authors: Diep Dao*, University of Colorado - Colorado Springs, Jean-Claude Thill, University of North Carolina - Charlotte
Topics: Spatial Analysis & Modeling, Geographic Information Science and Systems
Keywords: crime data mining, spatial association rule mining, SpatialARMED, urban residential motor vehicle thefts
Session Type: Paper
Day: 4/4/2019
Start / End Time: 5:00 PM / 6:40 PM
Room: Capitol Room, Omni, East
Presentation File: No File Uploaded

The field of crime data mining has been developed with known efforts in crime clustering and classification analysis. A limited number of works in this field examined the use of association rule mining to detect associations to crime or to motor vehicle thefts. While association rule mining is a promising technique to detect meaningful associations in spatial data generally, and in crime data particularly, important spatial components embedded in the studied phenomenon, e.g. autocorrelation, heterogeneity, and complex functional relations, need to be thoroughly considered. This paper tackles the challenge of enhancing spatial association rule mining (SARM) to consider autocorrelation and spatial interactions in urban motor vehicle theft analysis. The process of handling the spatial components during predication for SARM is presented. Spatial associations of both high and low urban motor vehicle thefts to neighborhood characteristics and to crime generators and attractors are revealed and discussed.

Abstract Information

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

To access contact information login