基于相关度的忙时话务量加权一阶局域预测模型

Weighting first-order local prediction model for busy hour traffic based on correlation

为了改善加权一阶局域预测模型的预测精度,提出衡量相点间相关性的相关度函数。利用相关度函数确定预测中心点的参考邻域,并将相关度的大小以加权的方式作用于加权一阶局域预测模型,进而推导出辨识加权一阶局域预测模型参数的算法。最后将预测模型用于忙时话务量的预测中,结果表明,该模型有效地提高了忙时话务量的预测精度,验证了相关度函数衡量相点间相关性的有效性。

In order to improve prediction accuracy of weighting first-order local prediction model, a correlation function is proposed to measure the correlation between different phase points, which is used to determine the referential neighborhood of central point. The correlation degree is acted on the weighting first order local prediction model in the form of weighting to de- rive the algorithm for identifying the parameters of weighting first-order local prediction model. The prediction model is applied to prediction of busy hour traffic. The test results show that the prediction model can effectively improve the prediction accuracy of busy hour traffic. The validity of correlation function to measure the correlation between different phase points was verified.