China
Africa
Explore chapters and articles related to this topic
Study on preparation of C4 alkenes by ethanol coupling
Published in Binoy K. Saikia, Advances in Applied Chemistry and Industrial Catalysis, 2022
Zhongzheng Wang, Xin Zhang, Yi Zheng, Hong Fang
To determine the result more accurately, we perform a White test on the OLS model and get that P = 0.7291 > 0.05, So we can reject the null hypothesis that there is heteroscedasticity.
Dynamic Capabilities, Eastern Relationships, and Competitive Advantages: An Empirical Assessment of Chinese and South Korean International Contractors
Published in Engineering Management Journal, 2022
Yi-Hsin Lin, Zheng Qin, Chan Joong Kim, Yujia Zhang, Nini Xia
This study constructs three models to test all the hypotheses. The Ramsey regression equation specification error test (Ramsey, 1969) is performed to test specification errors, such as omitted variables and the nonlinearity of functional form. The p-value of the said Ramsey test was .225 > .1, which indicated that there are no missing items. The White test is performed to examine the integrity of the sample’s homoskedasticity. The p-value of the White test for Model 1 was .120 > .1, Model 2 was .028 < .1, and Model 3 was .782 > .1, thus indicating that there are no obvious heteroskedasticity problems in Models 1 and 3, but there is such a problem in Model 2. Hence, the more stringent hetero-robust standard errors are adopted to determine the p-values of the regression coefficients.
The long-run and short-run effects of oil price on energy consumption in Tunisia: Evidence from structural breaks analysis
Published in Energy Sources, Part B: Economics, Planning, and Policy, 2020
We also test whether the estimated VAR model violates the assumptions of series correlation, constant variance, and normality. According to the results of the LM tests presented in Table 11, there is no serial correlation of the VAR model residuals up to lag 2, which makes it possible to accept the null hypothesis of the residuals absence of correlation. The Jarque-Bera normality test indicates that the VAR model residuals are normally distributed. The heteroscedasticity of the residuals is verified using the White test by including cross terms and without cross terms. The results show that there is no heteroscedasticity between the VAR model residuals in cross-terms and without cross-terms. Finally, the VAR model stability condition is tested by the autoregressive roots (AR). As a result, there are two real positive roots with two modules. The results indicate that the first root is outside the unit circle, which does not satisfy the stability condition of the VAR model.
Modeling and robust prediction of high heating values of municipal solid waste based on ultimate analysis
Published in Energy Sources, Part A: Recovery, Utilization, and Environmental Effects, 2020
Rotimi A. Ibikunle, Adewale F. Lukman, Isaac F. Titiladunayo, Emmanuel A. Akeju, Samuel O. Dahunsi
The diagnostic check shows that there are certain violations in the assumption of the linear regression model which makes the OLS estimator not the most efficient estimator for this modeling. The Jarque-Bera (JB) test shows that the error term is not normally distributed. The test statistic value is 48.2133 with a corresponding p-value of 0.0000. Thus, since the p-value is less than the level of significance (α = 0.05) then we conclude the error term is not normally distributed. The white test is employed to check if the error term has a constant variance. This shows there is no problem of heteroscedasticity. The white test value is 21.007 with a p-value of 0.7860. There is constant variance since the p-value is greater than the level of significance. The other problem exhibited in this model is the challenge of multicollinearity since the variance inflation factors for some variables are greater than 10. According to Gujarati (1995), there is multicollinearity when the VIF exceeds 10. The diagnostic test revealed that the model suffers the problem of non-normality and multicollinearity. For the purpose of handling these problems simultaneously, we employed the robust-ridge estimator, robust-Liu and robust-KL estimator. The performance of the estimators is compared using the scalar mean squared error (SMSE). The results are presented in Table 6 and the estimator with the least SMSE is considered best. The robust-KL estimator possesses the least SMSE. Figures 4 and 5 shows the graph of the predicted value against the actual value using both the OLS and the robust-KL estimator.