Machine Learning Based Prediction of Crime Rates in Districts in India

Abstract

Crime is a significant problem in India and around the world. Understanding trends in crime rates can help with understanding crime patterns in different districts. This project presents a system called CrimeSense which uses Machine Learning to predict crime rates in India based on historical crime data from different districts from 2014 to 2023. The dataset was cleaned and analysed using 4 different regression models; Linear Regression, Decision Tree, Random Forest and Gradient Boosting. The models were compared using several metrics including MAE, RMSE and R² score. Initial results show that Gradient Boosting and Linear Regression both provided similar results. Both having an R² score of 0.4898 and 0.4896 respectively.CrimeSense also provides ranked crime data for each district. The system was built using Python and Streamlit and makes crime data from previous years easier to analyse and view the results of the prediction. The aim of this project is to show that Machine Learning can be used to understand crime trends in different districts in India.