Comparison of Some Portmanteau Tests by Using Simulation

Abstract

Passing the chosen diagnosed candidate model is considered an employed modeled gate and the possibility of data presentation to be used in prediction and control. Also the research has dealt with different tests for Portmanteau and because the wide spread test in the ready made Programs is the L-B test, and sicne many studies mentioned that this test suffers from huge differences in variation this led us to deal with the most recent tests like H-M and L-M trying to study its trial empirical power and its statistical specification (mean, variance) and applying some modification through adopting the partial autocorrelation coefficient PACF instead of the autocorrelation coefficients ACF, when calculated and suggesting a new test depending on the mean of the autocorrelation coefficient -rk- that was suggest by Moran 1948 .To reach this target simulation technique was used to comparieson Portmanteau tests in addition to the suggested test .The study showed that the use of the PACF in portmanteau tests contributed to increase it's empirical power and at the same time decrease it's variance.