Forecasting Murder Motives in India Using Statistical Analysis and Explainable Artificial Intelligence

Forecasting murder crimes is an indispensable part of strategic planning in law enforcement efforts. The knowledge of the motives that influence murder crimes is no less significant for crime forecasting and prevention. In this paper, we propose a comprehensive multi-method approach that unifies advanced statistical tests, time series forecasting models, and Explainable Artificial Intelligence (XAI) techniques to predict murder counts and their motives in India. This study examines five decades of state-wise records, from 1972 to 2022, collected from NCRB. The research aims to test the significance of motives and forecast murder for each motive with the help of ARIMA, SARIMA, and LSTM with XAI interpretation, with sub-objectives including recommendations on actions for law enforcement and decision-makers. The LSTM model overpowers other models with an RMSE score of 0.6212 and an MAE score of 0.3289. With such an interdisciplinary method, the paper aims to add greatly to the progress of crime motive prediction and prevention techniques, hence enhancing public safety and societal resilience in India.