• Souglman Kabirou Kanlanfeyi
• Keshav Kishore

Since the evolution of AI techniques and their applications on in many domains such as banking, industries, administrations, agricultures... it has also been applied to on crime prediction or forecasting. Different studies showed that we can predict crime by using AI techniques such as machine learning classification algorithms to predict a type of crime by using time and location information. Since the deployment of different libraries designed for Deep Learning techniques like Tensorflow, Theanos and Keras, applications of DL on many fields has become very popular and so for crime prediction which leads to crime prediction by using DL which performs better than ML traditional algorithms. Because of the possible relationship between two or many occurrence of crime or in other words, considering that a crime occurrence can be related to previous ones, Recurrent Neural Networks (RNN) which is a DL architecture designed for time series problems has shown to perform better than not only ML but also than DNN algorithms. This paper reports different studies done in the domain of crime prediction using different AI techniques by considering different factors that can affect crime prediction and the limitation of some algorithms for predicting and their solutions such as RNN with LSTM.

Crime Prediction, Machine Learning, Deep Learning, RNN, LSTM

"A SURVEY OF CRIME PREDICTION USING MACHINE LEARNING AND RECURRENT NEURAL NETWORKS", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.6, Issue 6, page no.629-634, June 2019, Available :http://www.jetir.org/papers/JETIR1906H66.pdf

"A SURVEY OF CRIME PREDICTION USING MACHINE LEARNING AND RECURRENT NEURAL NETWORKS", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.6, Issue 6, page no. pp629-634, June 2019, Available at : http://www.jetir.org/papers/JETIR1906H66.pdf