基于灰色ELM神经网络的短时交通流量预测

Short-term traffic flow prediction based on grey ELM neural network

为了提高短时交通流预测的精度,针对现有灰色模型,利用一阶线性微分白化方程拟合交通流数据。针对交通流数据波动性较高和易失真的缺点,提出一种基于灰色ELM神经网络的短时交通流预测方法。首先对短时交通流数据采用灰色模型累加处理,将其转化为长时交通流数据,以降低交通流数据的随机性,有效减小因数据本身波动造成的误差。然后,利用ELM神经网络代替一阶线性微分白化方程,对长时交通流进行预测。最后,将长时交通流预测结果经过累减还原为短时交通流预测结果,有效提高了预测精度。仿真验证结果表明,相比于现有的一些预测方法,该方法提高了预测精度,是一种有效的短时交通流预测方法。

The existing grey models have the disadvantages of the higher fluctuation of data and distortion when using the first order linear differential whitening equation to fit traffic flow data.To improve the accuracy of short-term traffic flow prediction,a short-term traffic flow prediction model based on grey ELM neural network was proposed.Firstly,the short-term traffic flow data was accumulated by gray model to convert it into the long time traffic flow data and to reduce its randomness,which could effectively reduce the error caused by the fluctuation of the data.Then,ELM neural network was used to replace the first order linear differential whitening equation to predict the long-term traffic flow.Finally,the short-term traffic flow prediction results could be obtained by restoring the long-term traffic flow prediction results.And it improved the accuracy of prediction.The simulation results showed that the proposed method was an effective short-term traffic flow prediction method and was much more accurate than the existing methods.