ARMA(p,q)模型参数初估计新方法研究

来源期刊：中南大学学报(自然科学版)1998年第3期

论文作者：侯晓鸿 李一智

文章页码：99 - 101

关键词：ARMA(p,q)模型;参数初估计; Y-W法;逆函数法

Key words：Auto-Regressive Moving Average model; parametric preliminary esmitation; Y-W method; invertible function method

摘    要：提出了一种Y-W法与逆函数法相结合的ARMA(p,q)模型参数初估计新方法.该方法不但具有在计算机上容易实现的特点,而且与当前常用的参数初估计方法相比,在参数估计精度上有了较大的提高.这对ARMA(p,q)模型的进一步推广与应用将有着极为重要的意义.

Abstract: Based on Y-W method and invertible function method, a new parametric preliminary esmitation method of Auto-Regressive Moving Average model is presented. The new method has two main characteristics: one is that it can be easily implemented on computers, and the other is that it has greater improvement with respect to precision of parametric esmitation than any other one. So we think that the further application of the model will be of great significance.