Analysis of multivariate time series for some linear models by using multi-dimensional wavelet shrinkage

Abstract

The wavelet transform for multivariate time-series analysis and prediction is performed using an automatic slow distribution model and the prediction accuracy is compared with the normal method through the use of statistical criteria. By taking the variables represented by the gross domestic product as a response variable and air pollutants as explanatory variables represented by the emissions of nitrous oxide and carbon dioxide in Iraq. The study showed that there is an inverse relationship between the emissions of carbon dioxide and gross domestic product, while a positive relationship between the emissions of nitrous oxide and gross domestic product. And that the variables analysis and prediction after performing the wavelet transform of the data is the best because it contains the lowest values ​​of the mean square error criterion and the mean criterion of absolute error.