Scientific Journal of the National Academy of Internal Affairs

An innovative method of predicting crime described infrequently used form was developed by the University of California in Los Angeles (USA), which is made for theidentification and prognostic use of fractal regularities in random numerical ranks of criminal statistics. The «criminal induction» effectcaused by excitation of society as a result of criminal event lies at the heart of formation of these regularities. Estimated total rate of crime is determined by not only classical criminogenic reasons and conditions, but also by the «social self-excitation»effectas a reason of previously committed crimes. This method allows to localize the place and time of committing forthcoming crimes. Considering the results of applying this method by the Police Department of Los Angeles,the practice of police patrol daily instructing was implemented, appreciable reduction of street crime was achievedas a result.