基于混合EMD-BPNN方法的短期客流预测分析

Analysis of Short-term Passenger Flow Prediction Based on Hybrid EMD-BPNN Method

通过城市轨道交通的客流预测,可以达到提升乘客出行效率、降低运营成本等目的。基于此,提出一种经验模态分解和神经网络相结合的混合EMD-BPNN方法来预测短期的客流量。该方法通过经验模态分解将原始的客流数据分解成多个固有模态函数分量,并筛选出有意义的分量,将其作为神经网络的输入,从而进行客流预测。实验结果证明,该方法在地铁的短期客流预测中的精度和稳定性均高于传统神经网络算法。

Through the prediction of passenger fl ow of urban rail transit,the purpose of improving passenger travel efficiency and reducing operating costs can be achieved.Based on this,a hybrid EMD-BPNN method combining empirical mode decomposition and neural network is proposed to predict the short-term passenger fl ow.The method decomposes the original passenger fl ow data into multiple intrinsic modal function components through empirical mode decomposition,and fi lters out the meaningful components,which are used as the input of the neural network to predict the passenger fl ow.The experimental results showed that the method proposed in this paper is more accurate and stable than the traditional neural network algorithm in the short-term metro passenger fl ow prediction.