Mining regular crime patterns in spatio-temporal databases

Crime pattern theory emphasize how crime appear from Spatio-temporal method. To get knowledge on crimes, it has to be studied in terms of space and time. Crime patterns and crime related activities research those spatio (location) distribution of crimes, omitting the temporal (time) dimensions. Exactly separation of crimes by space and time, example hour of a specific day, day of a specific week, month of a specific year is suitable to theory proposed. Latest data enables spatio-temporal separation of data increasingly achievable; larger data files allow separation of larger data into time and space. Thus to identify the regularity of crimes with respect to time and space, we proposed a new algorithm called RCPSTD algorithm to mine (regularly crime patterns in spatio-temporal database) using a vertical data format which requires database scan of once. A crime pattern is said to be regular, whenever the occurrence behavior of a crime pattern is less than or equal to user given regularity threshold. Experimental results of RCPSTD algorithm is efficient in both memory utilization and execution time.