Gated Recurrent Unit (GRU) in RNN for traffic forecasting based on time-series data

Traffic management has emerged as the top priority in the advancement of smart cities, and traffic prediction has been a crucial area of study in smart transportation systems that emphasize minimizing road congestion. Many new concepts have emerged as a result of vehicle networks (VNs), like mapping of traffic, traffic management, and automobile communication. Machine Learning (ML) is an effective method for discovering hidden insights in ITS without explicitly programming it by learning from data. In this work, Gated Recurrent Unit (GRU), the method in Recurrent Neural Networks is used for the time series traffic analysis and prediction. When compared with previous models, this model has shown a considerable improvement in accuracy. The findings were computed using the widely used forecasting metrics MSE, MAE, and RMSE.