Crime prediction model with spatial decision support systems: The case of theft crime in Ankara province

In the dynamic structure of the 21st century, the fight against crime is in the interest of the whole society with its increasing importance. Among the requirements of the age and the basic duties of the law enforcement, preventing crime and predicting correctly beforehand, besides intervening in crime, took place. Crime maps, which have been actively used for the last 70 years, have gained different capabilities with the latest developments in geographic information systems (GIS). In this way, effective crime analyzes can be made and crime prevention tactics and strategies can be developed. Crime information systems, which make it easier to see the events in a single window and make quick decisions, are effectively used in crime analysis and crime prediction. In addition to methods such as Probable Hotspot (PHotspot) Method, Probable Kernel Density Estimation (PKDE) and Time-Space Scan Statistics (STSS) developed for location and time-based prediction capability of crimes, Spatial Decision Support Systems (SDSS) and other crime prediction and analysis with the Repeat and Near Repeat (RNR) Analysis method, an effective crime estimation can be made from the immediate environment. In the study, SDSS and RNR methods were used together, and it was concluded that the location-based factors affecting the formation of the crime, determined as a result of professional experience, are related to the repetition of the crime, which can be taken into account.