A Novel Learning Method for Traffic Flow Forecasting by Seasonal SVR with Chaotic Simulated Annealing Algorithm

Abstract

The prediction of traffic flow in cities has always been one of the most important issues in the study of road traffic congestion in the world. However, it is difficult to accurately predict the traffic flow between cities, because the traffic flow prediction process involves a more complex nonlinear data model, especially during the daily peak hours, the traffic flow data presents a cycle Sexual (seasonal) trends. In recent years, support vector regression (SVR) has been widely used to solve nonlinear regression and time series problems. This paper uses a combination of chaos theory and simulated annealing algorithm to optimize the kernel parameters of the correlation vector machine. However, for the time being, there is no recognized SVR model to deal with cyclical (seasonal) trend time series. This paper proposes a traffic flow prediction model, which combines seasonal support vector regression model and chaotic simulated annealing algorithm (SSVRCSA) to predict the traffic flow between cities. Under previous research, support vector regression using chaotic sequence and simulated annealing algorithm has shown its advantages, which can effectively avoid falling into local optimal. Experimental results show that the proposed SSVRCSA model can produce more accurate prediction results than other alternative methods. This research finally proposed a prediction model that blends the seasonal support vector regression model and the chaotic cloud simulated annealing algorithm (SSVRCCSA) to obtain more accurate prediction performance. The experimental results show that the proposed SSVRCCSA model is more accurate than other methods.