Traffic Flow Prediction Model Based on LSTM with Finnish Dataset

Abstract

Accurate prediction of traffic flow can achieve reliable traffic control and inducement. To solve the problems of complex traditional prediction models and insufficient prediction accuracy, this paper proposes a traffic flow prediction model based on long short-term memory (LSTM). First, a real traffic flow dataset is selected to macroscopically analyze the traffic flow from the lane level. After that, the training set and test set are divided, and the LSTM is used to predict the traffic flow. The results of this algorithm are compared with those of gated recurrent unit (GRU) and stacked autoencoders (SAEs), and the results show that this algorithm has the lowest traffic flow fitting error and the highest performance.