Gromadzenie danych i prognoza przestępczości z szeregów czasowych
Grzegorz Borowik , Zbigniew M. Wawrzyniak , Eliza Szczechla
AbstractTechnological development has shaped a broader analytical approach to crime forecasting. The increasing possibilities of tracking criminal events give public organizations and police departments the ability to collect and store detailed data, including spatial and temporal information. At the same time, open collections of social and Internet data may be a valuable source of knowledge about various patterns of behavior and social phenomena, including criminal ones. Thus, data analysis and mining become an important part of the current methodology for detecting and forecasting crime development. The ability to use data analysis tools allows effective deployment of officers in areas with high crime risk and elimination from areas with declining crime, as well as the development of effective crime prevention strategies. The purpose of this chapter is to demonstrate the usefulness of analytical algorithms for predicting crimes from time series. The most important factor is the accuracy with which one can deduce and create new knowledge based on past observations that will be useful in the process of reducing crime and ensure the safety of citizens. The chapter also presents issues and problems related to the process of collecting and pre-processing data. The purpose of numerical experiments based on observational data is the forecast for crimes in the period between July 1, 2016 and July 1, 2017. The research and analysis for police intervention, burglary, traffic offense, hooliganism, detention, and other criminal offenses were carried out. The experimental forecasting results obtained through the use of spectral analysis and Prophet package confirm the usefulness of the analysis. Weekly and annual seasonal patterns as well as the trend for selected event categories were identified. Forecasting models were developed for the analyzed time series of criminal events, which reached an average absolute percentage error between 10% and 40%, depending on the type of the event. While there is still considerable variation in crime that cannot be captured with predictive models, it is thought that these levels of forecasting quality are useful for a more efficient allocation of law enforcement forces.
|Publication size in sheets||2.15|
|Book||Hołyst Brunon, Malec Norbert, Wawrzyniak Zbigniew M.: Prognozowanie kryminologiczne w wymiarze społecznym, Tom 2 - Modele prognostyczne, Przestępczość, Wiktymizacja, Profilaktyka, 2018, Wydawnictwo Naukowe PWN, ISBN 978-83-01-20469-3, 1104 p.|
|Score||= 20.0, 20-10-2019, MonographChapterAuthor|
* presented citation count is obtained through Internet information analysis and it is close to the number calculated by the Publish or Perish system.