Crime Prediction and Forecasting using Voting Classifier

Abstract

The creation of policing strategies and the implementation of crime prevention and control programmes rely heavily on crime prediction. In India, the number of cognizable offences are gradually rising. These offences include those that are covered by the Indian Penal Code as well as those covered by a variety of Special and Local Laws. In this study, various machine learning algorithms are used to analyze data related to crime which will give the behaviors in crime over an area, that might be helpful in crime prevention. Classification Algorithms such as KNN, Random Forest, Adaptive Boosting Classifier, Gradient Boosting Classifier, and Extra Trees Classifier were tested individually on a crime dataset based in South Bangalore. Later, these five algorithms were combined to achieve better results using Voting Classifier Technique in order to balance out their individual weaknesses.