Time Series Analysis: Regression Techniques, Том 9;Том 1990

ACF and PACF Applying OLS approach AREG assume autocorrelation function Beach-McKinnon bias Box-Jenkins coefficient constant variance correlograms dependent variable discussed distributed lag disturbance term e₁ EGLS estimates EGLS residuals equation 2.1 error process error term esti estimated residuals evaluation ex-ante-forecasts ex-post-forecasts example exogenous expected value Figure first-order autoregressive process following command forecast error Hibbs Hildreth-Lu instrumental variables Johnston Kmenta lagged endogenous variables lagged values Malinvaud McCleary and Hay moving average process naive model nonautoregression assumption Note null hypothesis numbers in parentheses observations obtain OLS estimates Ostrom P₁ parameters positive serial correlation possible postsample Prais-Winsten present problem procedure random random variable ratio goal model revised ratio goal sample serial correlation series regression analysis series regression model SPSS t-ratios technique timated time-dependent process tion transformation U.S. defense expenditure U.S. defense spending U.S. spending USSR USSR spending V₁ Wonnacott and Wonnacott X₁ Y₁ σ²