摘要
本文提出一种求取MA模型参数的线性迭代算法,它利用预测模型对观测数据进行白化处理,然后再通过反卷积方法求取模型参数.本文导出了预测反卷积法(PDC)的多步迭代形式,以改进参数估计.统计分析和实验结果表明,预测反卷积法是渐近无偏,一致的算法,具有简单易求的特点.
A linear algorithm for estimation of the parameters of moving-average (MA) models is presented. The algorithm uses a predictive model that makes whitening treatment of the obsevations and a predictive deconvolution (PDC) method to estimate the MA parameters.An iterative procedure of the PDC method is derived. The statistic analysis and experiments show that the PDC estimator is asymptotically unbiased, consistent, and has simple and convenient features.
出处
《信号处理》
CSCD
北大核心
1991年第3期173-178,共6页
Journal of Signal Processing
