ANN Based Crime Detection and Prediction using Twitter Posts and Weather Data

Crime has grown tremendously because of a growing population. So, crime prediction is an important instrument for the safety of civilians to aid law enforcement officials. Crime prediction can be done by analyzing historical data. But, it is a challenging task due to the unavailability of data. Further, some crimes are unregistered. Currently, people are using social media like twitter to casual chats, sharing photos and ideas, and transferring information and news. Thus, Social networking sites can be used to uncover useful information. In this paper, we propose a crime prediction approach using twitter data and weather data. The proposed approach contains two modules namely crime detection and crime prediction module. The crime detection module is used to detect the crime-related twitter posts. Bidirectional Encoder Representations from Transformers (BERT) based approach is used to develop the detection module. Then, the prediction module is implemented using an Artificial Neural Network (ANN). The empirical study of our proposed approach has proved the effectiveness of the BERT and ANN-based prediction approach.