Title page for ETD etd-1103103-095701


Type of Document Master's Thesis
Author Kent, Joshua David
URN etd-1103103-095701
Title Using Functional Distance Measures When Calibrating Journey-to-Crime Distance Decay Algorithms
Degree Master of Natural Sciences (M.N.S.)
Department Natural Sciences (Interdepartmental Program)
Advisory Committee
Advisor Name Title
Michael Leitner Committee Chair
Andrew Curtis Committee Member
DeWitt Braud Committee Member
Keywords
  • euclidean distances
  • functional distances
  • journey-to-crime
  • geographic profiling
  • serial crime
  • crime mapping
Date of Defense 2003-10-17
Availability unrestricted
Abstract
Spatial analysis has long been a valuable tool used within the criminal investigative process. This is especially true for serial offence cases where criminologists apply geographic profiling to model offender mobility and crime distribution patterns in order to estimate a criminal’s likely residence. Yet, traditional analytical methodologies have avoided the utilization of functional distance measures when modeling an offender’s journey-to-crime within an anisotropic landscape. By substituting straight-line Euclidean distances with travel path functional distance measures, the predictive utility and technological prerequisites associated with geographically profiling a localized serial offender was assessed using mathematically calibrated distance decay models. Both the travel-path and temporally optimized functional distance measures were calculated based on the impedance attributes stored within a linearly referenced transportation data layer of East Baton Rouge Parish, Louisiana. Journey-to-crime distance decay algorithms were mathematically calibrated for ‘best fit,’ based on the distribution of incidents obtained from a calibration sample of thirty-one simulated offenders. Functional distances measured for a series of incident locations attributed to a sample of four simulated serial offenders. Using the calibrated distance decay function measured from the calibration sample, geographic profiles were created for each of the four simulated serial offenders. A probability score was calculated for every point within the study area to indicate the likelihood that it contained the offender’s residence. Score surfaces estimating the likely residence of the sample offenders were calculated and compared to the actual, known residences in order to determine predictive value and procedural validity of functional distance metrics. Analyzed results revealed that the functional distance measures can serve as a substitute for traditional Euclidian distances when estimating the likely residence of a localized serial offender.
Files
  Filename       Size       Approximate Download Time (Hours:Minutes:Seconds) 
 
 28.8 Modem   56K Modem   ISDN (64 Kb)   ISDN (128 Kb)   Higher-speed Access 
  Kent_thesis.pdf 1.26 Mb 00:05:50 00:03:00 00:02:37 00:01:18 00:00:06

Browse All Available ETDs by ( Author | Department )

If you have questions or technical problems, please Contact LSU-ETD Support.