This study aims to examine the problem of autocorrelation, simple regression analysis with errors following the form of first-order autoregressive, Durbin Watson method using the Cochrane-Orcutt, Hildreth-Lu procedures and first differences in overcoming autocorrelation. The occurrence of autocorrelation causes the alleged regression parameter with ordinary least square (OLS) not to produce the actual value. Therefore, to obtain the actual parameters applied Durbin watson method with all three procedures. Based on the data used in this thesis quoted from the book Applied Linear Statistical Models Fifth Edition, the best procedure is given by Hildreth-Lu because it produces the smallest mean square error (MSE) value. This is because, the process of estimating the autocorrelation coefficient is based on iterations until a minimum sum square error (SSE) value is found.
Copyrights © 2023