Question

AaycGAaçcGgi Aaççãį Aaççõõt Aaç 1 Normal 1 Aralık Yok Başlık 1 Baylik 2 Konu Bayle Paragraf Stiller . State whether the following statements are true or false. Give your reasoning. Statistical inference in NLLS regression cannot be made n the basis of the usual t, F, and X2 tests even if the error rm is assumed to be normally distributed. The coefficient of determination (R2) is not a particularly eaningful number for an NLRM.

help me pls thank you so much

Answer 1

Answer a. True

Explanation:

Even if error terms are normally distributed, the Non-Linear-Least-Square (NLLS) estimators are not normally distributed. Thus, they are not unbiased and do not possess minimum variance for large as well as for small samples. Therefore, we cannot use t-test (for testing significance of individual coefficients), F-Test (to test overall significance of regression model) as we are unable to obtain unbiased estimator of error variance from estimated model and Ch-Square Test.

Answer b. True

Explanation:

The Difference between actual Y values and Fitted Y Values from Non Linear Regression Model i.e. the residuals from NLRM may not necessarily sum to 0 and the addition of Sum of Squares due to Error and Sum of Squares due to Regression may not necessarily equal Total Sum of Squares so,

$\mathbf{R^2 = \frac{SS.due.to.Regression}{Total.Sum.of.Squares}} = \frac{SSR}{TSS}$ may not be meaningful.