周相似特性下交通流组合预测方法研究

根据交通流量具有周相似的特性,构造了周相似序列。用霍特指数平滑法对周相似序列进行预测,用人工神经网络对残差部分进行预测。将指数平滑法与神经网络法相结合,以便发挥每种方法的优势,获得比单个方法更好的预测结果。实例分析表明,比单独使用ARIMA或单独使用神经网络方法,使用组合方法的预测误差最小,适合于实时的交通流预测。

Study on hybrid approach for traffic flow prediction based on weekly similarity

Weekly similarity time series are constructed based on the weekly similarity of traffic flow. The Holt's Exponential Smoothing(ES) method is employed to produce the forecasts for the weekly similarity time series,and Artificial Neural Network (ANN) method is used to produce the forecasts for the residual time series.The hybrid approach combining both ES and ANN makes use of the advantages of each method,so as to produce better prediction than that from single method.Experimental results with real data sets indicate that the combined method can produce more accurate predictions than that from ARIMA model or ANN alone.The hybrid model can be used for real-time short-term traffic flow forecasting.