Leveraging Gradient Boosting for Crime Prediction: A Focus on Women's Safety in India

In India, crimes against women is a concerning issue where incidents of rape, domestic violence etc are increasing at a very alarming rate. So predicting crime rates especially among women has become very critical. The data that is gathered from various crime incidents can be collected and preprocessed to further train models which can predict crime rates in advance. In this research paper, we used Gradient Boosting Regressor to analyze crime data. This data is further used to train our machine learning model. The purpose of our study is to evaluate the Gradient Boosting Regressor's effectiveness and performance in predicting different kinds of criminal incidents. Our model predicts a range of criminal activities, such as sexual assault, deaths related to dowries, abuse by spouses or their family members, kidnapping and abduction, insulting women's modesty, and the importation of girls. Our model demonstrated high predictive performance, with R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> values above 0.99 for most crime categories, indicating its ability to explain the majority of the variance in crime rates effectively. The model's robustness was confirmed by low MAE and MSE values, along with consistent cross-validated MSE, suggesting its generalizability to unseen data. The findings show how advanced algorithms based on ML can be used to analyse crime data and provide a reliable method for identifying and predicting patterns in criminal activity.