Prediktivní analýza vloupání do objektů na území Ostravy

Abstract

The thesis is focused on whether it is possible to predict the crimes (in the case of this thesis burglary) in the city of Ostrava. The data was provided by the Czech Police for the years 2010 - 2011. This data was prepared in advance (mainly geocoded) from earlier student works. In the first part, emphasis is placed on the burglary, the use of burglary mapping methods and the prediction possibilities. The first practical part was devoted to exploratory data analysis (statistical and spatial). In the second part of the practical work, burglary mapping was performed in individual time periods. The last part was devoted to the actual evaluation of map outputs if they are suitable for prediction. This was done using the hit rate and PAI index methods. Finally, the benefits and recommendations for further work were evaluated.