Abstract
We introduce a method that eliminates the specification error and spurious relationships in regression. In addition, we introduce a test of strong causality. Furthermore, hypothesis testing (inference) becomes almost unneeded. Moreover, this method virtually resolves error problems such as heteroscedasticity, autocorrelation, non-stationarity and endogeneity.
Language: English
Page range: 45 - 48
Published on: Jun 22, 2018
Published by: Polish Biometric Society
In partnership with: Paradigm Publishing Services
Publication frequency: 2 issues per year
Related subjects:
© 2018 Moawia Alghalith, published by Polish Biometric Society
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.