Nonparametric regression model of autocorrelation problem in the general budget of Arab States through (2008-2020)

摘要

In this research were addressed estimators nonparametric to nonparametric in case of a problem autocorrelation of by editing the least squares method based on support vector regression (LS-SVM) using standard intersection of leave - one-out (Cross - Validation) debugger to correlation error (CC-CV) and eat estimated medullary for Nadaraya Watson with the correlation error by addressing the estimated algorithm medullary for Nadaraya Watson with the correlated error using standard the intersection of legitimacy of leave - one-out (Cross - Validation) correlated error (CC-CV) (NW) , eating (method proposed) estimated medullary for weighted Nadaraya Watson with the correlated error and that by eating medullary for weighted Nadaraya Watson estimated algorithm with the error correlation using standard intersection of leave - one-out (Cross - Validation) to correlated error (CC-CV) (RNW), , through the application of these methods on realistic data indicate the general budget for the Arab States for the period (2008-2020) , the results have proved destined estimated medullary preference for Nadaraya Watson-weighted with the correlated error using standard intersection of leave - one-out (Cross - Validation) to correlated error (CC-CV) (R.N.W) for the model by receiving a lower value to standards (RAMSE), (AMAPE).