PROBABILISTIC FORECASTING OF TRAFFIC FLOW USING KERNEL BASED EXTREME LEARNING MACHING AND QUANTUM-BAHAVED PARTICLE SWARM OPTIMIZATION

Accurate and reliable forecast of traffic flow is important to intelligent transportation systems (ITS). However, due to the non-stationary of traffic flow data, traditional point forecasting can hardly be accurate, so probabilistic forecasting methods are essential for quantification of the potential uncertainties and risks for traffic management. This paper proposes a kernel extreme learning machine (KELM)-based probabilistic forecasting method of traffic flow, and the parameters of KELM are optimized by using Quantum-behaved particle swarm optimization (QPSO) algorithm. To verify its effectiveness, traffic flow prediction using QPSO-KELM are compared with other methods. Experimental results show that QPSO-KELM has higher prediction accuracy at different confidence levels for practical applications in traffic management systems. And it will help traffic managers to make right decisions.