时间序列AR模型参数的积分求解法

An Integration Method for Solving the Time Series AR Model Parameters

基于对时间序列实质的分析，提出了旨在减少序列的随机误差影响以及提高拟合精度的ＡＲ模型参数的积分求解法．重点讨论了ＡＲ（１）模型及ＡＲ（ｐ）模型参数的积分求解法，并与最小二乘法在计算机上进行了仿真比较．结果表明，采用积分求解法所得的ＡＲ模型参数的估计精度比最小二乘法的高．

Based on an analysis of the crux of time series,an integration method for solving thetime series AR model parameters aimed at reducing the effect of the random error and im-proving the fitting accuracy is developed.Attention is focused on the solution of parametersof the AR(1) model and AR(p)model.The results obtained are compared with those by theleast square method through computer simulation. It is shown that the integration method isbetter than the least square method in terms of the estimation accuracy.