Crime Prediction on Open Data in India Using Data Mining Techniques

The growth in crime data collection necessitates a deeper theoretical understanding to support practical crime prevention strategies tailored to specific locations and times. Models for forecasting the frequency of various sorts of crimes under an administrative structure of regions utilized by the Indian police, as well as the frequency of anti-social behavior offenses, are investigated in this study. Four algorithms from several techniques are employed. The information is sourced from the Indian police and encompasses over 200,000 records before preprocessing. Based on predictive performance and processing time, the results indicate that SVM may be utilized to forecast crime frequency.Additionally, an ensemble of data mining classification approaches is employed to anticipate crime. Finally, the optimal forecasting technique for achieving the most stable results is presented. The study yields a model that utilizes implicit and explicit data mining approaches to produce credible crime predictions.