Traffic Flow Prediction Based on Long Short Term Memory Network

摘要

This study proposes a traffic flow prediction method based on long short term memory (LSTM) network. Firstly, traffic date is preprocessed by time series method. Then a traffic flow prediction algorithm framework based on LSTM arm was proposed to improve the accuracy of traffic forecast and compare algorithm differences between LSTM, support vector machine (SVM) and radial basis function (RBF). In the last part, a reliable experiment was designed. The experimental results verify the superiority performance of LSTM over SVM and RBF in traffic flow prediction.