On Diagnostic Checking Time Series Models with Portmanteau Test Statistics Based on Generalized Inverses and

摘要: A class of univariate time series models is considered, which allows general specifications for the conditional mean and variance functions. After deriving asymptotic distributions residual autocorrelations based on standardized residuals, portmanteau test statistics are studied. If covariance a vector fixed length non singular, could be defined, following approach advocated by Li (1992). However, assuming invertibility may restrictive, and, alternatively, popular Box-Pierce-Ljung statistic recommended. In our framework, that converges in distribution to weighted sum chi-square variables, critical values can found using Imhof’s (1961) algorithm. algorithm consuming. view this, we investigate this article use generalized inverses {2}-inverses, order propose new with distributions, avoiding need implement small simulation study, compared: statistic, proposal (1992), relying {2}-inverses.