| |
| Title |
An Evaluation of Spatial Forecasting Techniques Used in Tactical Crime Analysis
|
| Author(s) |
Van Auken, John B.
|
| School/Department |
Department of Geography
|
| Institution |
University of Denver
|
| Degree Type |
Master's
|
| Degree Name |
M.A.
|
| Type of Resource |
text
|
| Degree Date |
06/2005
|
| Digital Origin |
reformatted digital
|
| Rights Statement |
All Rights Reserved
|
| Reason for Restrictions |
No restrictions
|
| Type of Restriction |
No restrictions
|
| Keyword(s) |
Geography
Geology
|
| Genre |
Dissertations, Academic
|
| Abstract |
Spatial forecasting is used in tactical crime analysis to "predict" the area of a future event in a crime series. This allows law enforcement to take a proactive approach to criminal activity instead of a reactive approach. The concept has received recent publicity and scrutiny as a valuable new option in law enforcement due to various successes across the country. A survey of the crime analysis community on the use of spatial forecasting techniques was conducted to determine the most popular techniques in use today. The Minimum Convex Polygon, 68% and 95% Standard Deviation Rectangles, 50% and 90% Jennrich-Turner Ellipses, and Kernel Smoothing based on Spider Distance Analysis were determined to be the more prevalent techniques. These methods were looked at in depth for their ability to "forecast" the area within which the next event in a crime series will occur by evaluating them on thirty-one solved crime series. The series were collected from various analysts across the country, have a varying number of events and geographic distribution, and are of different crime types. The techniques were evaluated on the number of correct predictions and size of forecasted areas. The history and evolution of these techniques towards use in crime analysis is discussed, along with the possibilities for future research.
|
| Handle |
http://hdl.handle.net/10176/codu:55688
|
|
|
|
|
