LAC准则及其在ARMA模型辨识中的应用

LAC Criterion and its Application to ARMA(p,q) Model Identification

基于残差序列最小自相关原理，提出了模型适用性ＬＡＣ检验准则及其递推算法，用于辩识和估计ＡＲＭＡ（ｐ，ｑ）模型的结构和参数，在实际应用中取得了较满意的结果。其特点是计算量小、精度高，并且不受ｑ＜ｐ条件的限制。

Based on the least-autocorrelation of the residual sequence, the LAC testing criterion of model fitting and its recursive algorithm are proposed to identify and estimate the structure and parameters of ARMA (p,q) models. Many successful results were obtained in practice. This algorithm is characterized by its less computation, high accuracy and no restriction of the condition q<p.