Short term traffic flow prediction based on improved LSTM model

Abstract

Aiming at the problem of traffic flow prediction in intelligent transportation system (ITS), an improved long-term and short-term memory network (LSTM) model is proposed. This model uses attention mechanism to give different weights to the characteristics of different time points in historical time series, and then it can mine more effective information and improve the accuracy of the model when predicting future traffic flow. By comparing with support vector regression (SVR), BP neural network and LSTM, it can be found that this model has more effective prediction accuracy and model stability.