Portal de Periódicos da CAPES

Multiple sources and multiple measures based traffic flow prediction using the chaos theory and support vector regression method

2016; Elsevier BV; Volume: 466; Linguagem: Inglês

10.1016/j.physa.2016.09.041

1873-2119

Anyu Cheng, Xiao Jiang, Yongfu Li, Chao Zhang, Hao Zhu,

Complex Systems and Time Series Analysis

This study proposes a multiple sources and multiple measures based traffic flow prediction algorithm using the chaos theory and support vector regression method. In particular, first, the chaotic characteristics of traffic flow associated with the speed, occupancy, and flow are identified using the maximum Lyapunov exponent. Then, the phase space of multiple measures chaotic time series are reconstructed based on the phase space reconstruction theory and fused into a same multi-dimensional phase space using the Bayesian estimation theory. In addition, the support vector regression (SVR) model is designed to predict the traffic flow. Numerical experiments are performed using the data from multiple sources. The results show that, compared with the single measure, the proposed method has better performance for the short-term traffic flow prediction in terms of the accuracy and timeliness.