基于小波变换遗传过程神经元网络的交通流预测

针对短时交通流时间序列的缺点,应用小波变换理论,将含有综合信息的时间序列分离为低频确定信号和高频干扰信号,用遗传过程神经元网络分别进行预测,得到了原时间序列的实际预测结果,通过实测数据验证表明,该预测方法具有较好的预测精度。

Traffic flow prediction based on wavelet transformation epigenetic process neural networks

In light of defects of short-time traffic flow time series, the paper applies wavelet transformation theory, divides eomprehensive time series into low-frequency determination signal and high-frequency disturbing signal, carries out a prediction by using epigenetic process neural networks, and obtains the actual prediction results of original time series. As a result, the actual testing data proves that, the prediction method has better prediction accuracy.