非等间距相关系数平稳序列分析方法

Analysis Method for Correlation Coefficient Stationary Series with Unequally Spaced Data

工程实际中由于数据缺损或受测试条件所限,常常得到非等间距的时间序列。而传统的时序分析方法只适用于等间距采样的数据,目前工程上对这类问题通常采用插值等近似处理方法,这往往导致较大误差。本文提出非等间距相关系数平稳序列的概念,建立了非等间距相关系数平稳序列自回归模型,给出了非等间距相关系数AR(p)序列的精确极大似然估计和条件极大似然估计,能够高精度地确定非等间距序列的均值函数、方差函数和相关系数函数。

The idea of correlation coefficient stationary series with unequally spaced data is presented and the pthorder autoregressive model for the correlation coefficient stationary series with unequally spaced data is established.The exact maximum likelihood estimations and the conditional maximum likelihood estimations of their parameters are given in detail.Unequally spaced time series is common in engineering when the observation is irregular. The traditional time series analysis method is only suitable to equally spaced data and the approximate analysis methods for unequally spaced data, such as interpolation method,may cause big error.By employing the present method, the mean function,the variance function and the correlation coefficient function of unequally spaced time series can be calculated with high precision.