Creation of an IT system for the Polish Police supporting the short-term prediction of events, including crimes and offenses, suggesting to officers the place and time of a possible event.
The Polish Police is a centralized, armed and uniformed formation. Nearly 100,000 policemen supported by almost 25,000 civilian workers are expected to watch over the safety of Poland’s inhabitants and the maintenance of public order. The contemporary Polish Police consist of officers employed in the criminal, preventive and support services of the police in the organizational, logistic and technical areas.
Well-armed and trained anti-terrorist police are involved in detaining the most-dangerous criminals. Police officers of the Central Bureau of Investigation (CBŚP) unit break up organized crime groups, fight against criminal terror and drug trafficking.
ML learning methods are very effective in predicting criminal events. They can be used, among others to predict the place and time where such events may occur and to find spacetime relationships.
Predictive system assumptions: The MIM Solutions solution is based on the assumptions of the PredPol system. The PredPol system is the first crime prediction system to achieve significant success in crime prevention. It operates in 11 major cities in the US and the UK. The main advantage of this system is the very small amount of data needed to make forecasts and thus the speed and flexibility of implementation in various centers. The basic system is based solely on the place, time and type of event. The MIM Solutions system has a supporting task and is to suggest to officers the place and time of a possible event. The system extracts various crime patterns from historical data (e.g. on Fridays in the evening there are more robberies in this district, on working days there are break-ins most often in this place, etc.) that a pair of human eyes may not notice.
Data used to create the model: The predictive system was based on police data from the decision support system (DSS, in Polish: SWD), which includes the place, time and type of event. This data gets processed to obtain the exact coordinates of the place based on the description.
Creating a predictive system: The MIM Solutions model uses a spatial-temporal regression model based on algorithms, such as gradient boosted decision trees (GBDT). It predicts the probability of another event based on previously observed events in the area. It allows the simple integration of external data sources such as weather, cultural and sports events. It allows you to make predictions for any period of time, e.g. hour, day, month. When preparing a crime forecast for a given region (e.g. city, commune, county), we divided the region into small square areas called locations. The default and recommended square size is 200 by 200 meters. The features that the algorithm uses are a key aspect of any machine-learning algorithm. We made sure that the features only include information about past crimes – the information that will be available when preparing the forecast.
Results are based on the example of Białystok, a city in eastern Poland. With 20 police patrols at our disposal, we can immediately place a patrol at the scene of an offense in over 30 percent of cases. As the number of policemen increases, so does the number of successful preventive actions.