论文部分内容阅读
本文在分析了建立AR模型的Y-W矩估计算法原理后,针对其存在的不足,提出了修正的矩估计法。该方法对数据的自相关系数进行最小二乘拟合,使估计的模型对谱有很好的拟合性。本文推导了两种逐阶递推算法且通过计算机模拟实验说明修正矩估计法的精度优于Y-W法。
After analyzing the principle of Y-W moment estimation algorithm for establishing AR model, aiming at its shortcomings, this paper proposes a revised moment estimation method. This method performs the least-squares fitting on the autocorrelation coefficient of the data and makes the estimated model fit well to the spectrum. In this paper, two kinds of step-by-step recursive algorithms are deduced and computer simulation results show that the accuracy of the revised moment estimation method is better than the Y-W method.