Long short-term memory (LSTM) recurrent neural network (RNN) based traffic forecasting for intelligent transportation

Abstract

The Long short-term memory is one of the scheme in Deep Learning technique that can be used for more accurate time series forecasting. Cities are nowadays facing a huge traffic congestion due to the rapid increase in the number of vehicles. More accurate traffic forecasting is very much essential for an Intelligent Transportation System (ITS). The advancement in the Computational Intelligence (CI) techniques provide best results in the forecasting problems compared to the classical statistical approach. The proposed work is based on the CI technique, the LSTM for the traffic flow forecasting using traffic sensor data that are available as a big data. The experiment shows better forecast metrics with RMSE (Root Mean Squared Error) of 8.4 and 9.86 for training and test data respectively. The MAE (Mean Absolute Error) is 5.52 and 6.83 for training and test data respectively. Thus the proposed work on LSTM based traffic flow prediction results in a superior performance.