Short-Term Traffic Flow Prediction Model Based on BP Neural Network Algorithm

Abstract

Nowadays, the application of traffic data collection in major cities around the world is constantly updated, which promotes the continuous improvement of short-term traffic prediction capabilities. The purpose of this paper is to study the short-term traffic flow (TF) prediction model based on BP neural network (NN) algorithm. This paper presents a new forecasting model and an algorithm for forecasting short-term TF. Short-term traffic forecasting will play a very important role in traffic management applications. This paper proposes a prediction algorithm based on BP NN, and uses the optimal density rule to improve the prediction algorithm. Based on the demand structure of the TF forecasting system, this article outlines the overall program flow. This article conducts TF forecasting analysis and processing on the experimental data obtained and used by the BP network, and compares them through experiments. It is concluded that the method is effective in improving TF. The accuracy of the forecast is valid. Experimental research shows that the degree of fit between the test output curve and the expected output curve in this paper is enhanced. In the case of relatively large actual output fluctuations, after GA optimizes the weights and thresholds of the BP NN, the convergence rate of the predicted output curve and the prediction accuracy are increased by 20%.