Urban traffic flow forecasting using Gauss–SVR with cat mapping, cloud model and PSO hybrid algorithm

In order to improve forecasting accuracy of urban traffic flow, this paper applies support vector regression (SVR) model with Gauss loss function (namely Gauss–SVR) to forecast urban traffic flow. By using the input historical flow data as the validation data, the Gauss–SVR model is dedicated to reduce the random error of the traffic flow data sequence. The chaotic cloud particle swarm optimization algorithm (CCPSO) is then proposed, based on cat chaotic mapping and cloud model, to optimize the hyper parameters of the Gauss–SVR model. Finally, the Gauss–SVR model with CCPSO is established to conduct the urban traffic flow forecasting. Numerical example results have proved that the proposed model has received better forecasting performance compared to existing alternative models. Thus, the proposed model has the feasibility and the availability in urban traffic flow forecasting fields.