Leveraging Machine Learning and Natural Language Processing for Predicting the Crime Rate: Reach 360
Proceedings of 3rd International Conference on Internet of Things and Connected Technologies (ICIoTCT), 2018 held at Malaviya National Institute of Technology, Jaipur (India) on March 26-27, 2018
9 Pages Posted: 8 May 2018
Date Written: April 29, 2018
Abstract
As the magnitude of the crimes being reported is rapidly rising, ensuring the safety of women becomes an utmost concern in India. Identifying trends and patterns in massive volume of data available on crime becomes paramount, but the enormity of the task makes it time consuming and labour intensive for the human checkers. Thus, in wake of the current scenario, we were motivated to develop a comprehensive safety solution-Reach 360 an android application that utilizes the open data provided by National Crime Records Bureau(NCRB) along with online news sources to deploy machine learning models to accurately predict crime in the future. Our neural network model-multi layer perceptron provided an accuracy of 99.82% outperforming Random Forest and Logistic Regression. Apart from this, heatmap visualization of crime hotspots and shake alert features in the application serve as a perfect solution for ensuring safety at all times.
Suggested Citation: Suggested Citation