An Implementation of Traffic Volume Forecasting System using Distributed Deep Learning

The increase of means of transport, on the one hand, makes traveling easy and convenient, and unintentionally makes traffic congestion more and more serious, especially in big cities. Traffic congestion has a direct and indirect impact on a country's economy and its dwellers' health. Traffic congestion effects on individual level as well. Time loss, especially during peak hours, mental stress, and the added pollution to the global warming are also some important factors caused due to traffic congestion. In recent years, traffic volume prediction has received lots of interest from researchers. However, there are still many challenges that needs to be resolved. Models for traffic volume prediction needs not only to be accurate but also be able to handle the huge volume of input traffic data. This study present an implement of a traffic volume forecasting system using distributed deep learning. A state of the art deep learning model, that is LSTM, is used to provide accurate forecasting result. Especially, the model is trained in a distributed manner using BigDL library. Thank to the distributed training, the model now is able to receive a huge volume of traffic data to provide the most accurate results.