Clustering residential burglaries using multiple heterogeneous variables
(English)In: International Journal of Information Technology & Decision MakingArticle in journal (Refereed) Accepted
To identify series of residential burglaries, detecting linked crimes performed bythe same constellations of criminals is necessary. Comparison of crime reports today isdicult as crime reports traditionally have been written as unstructured text and oftenlack a common information-basis. Based on a novel process for collecting structured crimescene information the present study investigates the use of clustering algorithms to groupsimilar crime reports based on combined crime characteristics from the structured form.Clustering quality is measured using Connectivity and Silhouette index, stability usingJaccard index, and accuracy is measured using Rand index and a Series Rand index.The performance of clustering using combined characteristics was compared with spatialcharacteristic. The results suggest that the combined characteristics perform better orsimilar to the spatial characteristic. In terms of practical signicance, the presentedclustering approach is capable of clustering cases using a broader decision basis.
Crime Clustering; Residential Burglary Analysis; Decision Support System; Combined Distance Metric
IdentifiersURN: urn:nbn:se:bth-10807OAI: oai:DiVA.org:bth-10807DiVA: diva2:861200