Regression.pdf

# Note that the anova sums of squares are all quadratic

• 63

This preview shows pages 14–17. Sign up to view the full content.

Note that the ANOVA sums of squares are all quadratic forms. i) SST = Y Y 1 n Y 11 Y = Y [ I 1 n 11 ] Y ii) SSR = ˆ β X Y 1 n Y 11 Y = Y [ H 1 n 11 ] Y iii) SSE = Y Y ˆ β X Y = Y [ I H ] Y PAGE 14

This preview has intentionally blurred sections. Sign up to view the full version.

1.3 Matrix approach to Simple Linear Models c circlecopyrt HYON-JUNG KIM, 2017 We can also test a hypothesis about a linear combination of parameters, l β . l ˆ β Normal ( l β , l (X X) 1 l σ 2 ) . A two sided test: H 0 : l β = M vs. H a : l β negationslash = M can be done as follows. Reject H 0 if | t | = vextendsingle vextendsingle vextendsingle vextendsingle l ˆ β M vextendsingle vextendsingle vextendsingle vextendsingle s ( l ˆ β ) > t 1 α 2 ,n p , s( l ˆ β ) = radicalbig l (X X) 1 l MSE. (Recap) Under the normality assumption, the 100(1 α )% confidence interval for β k is ˆ β k ± t 1 α 2 s( ˆ β k ) , k = 1 , 2 s( ˆ β k ) = radicalBig (X X) 1 kk MSE . Simple example. DATA (X,Y): (4,5), (7,7), (5,6), (4,6) Assume that σ 2 = 0 . 25 known. 5 7 6 6 = 1 4 1 7 1 5 1 4 bracketleftBigg β 0 β 1 bracketrightBigg + ǫ 1 ǫ 2 . . . ǫ n We find the estimated parameters using the formula ˆ β = (X X) 1 X Y = (3 . 5 , 0 . 5) (Verify!) The variance-covariance matrix of parameter estimates is bracketleftBigg 4 . 4167 0 . 8333 0 . 8333 0 . 1667 bracketrightBigg (0 . 25) = bracketleftBigg 1 . 1042 0 . 2083 0 . 2083 0 . 0417 bracketrightBigg (1) Test that the slope is 0: t= 0 . 5 / 0 . 0417 = 0 . 5 / 0 . 2041 = 2 . 45 (2 degrees of freedom) (2) Test that the intercept is 0: (2 degrees of freedom) (3) Estimate the mean value of Y at X = 6 and give a 95% confidence interval. We are estimating: β 0 + 6 β 1 = [1 , 6] bracketleftBigg β 0 β 1 bracketrightBigg and l ˆ β = hatwidest Var( l ˆ β ) = l hatwidest Var( ˆ β ) l = [1 , 6] bracketleftBigg 1 . 1042 0 . 2083 0 . 2083 0 . 0417 bracketrightBigg bracketleftBigg 1 6 bracketrightBigg = 0 . 1042 . The 95% confidence interval for our little example is Note that the 97.5% quantile of the t distribution with 2 d.f. = 4 . 30. PAGE 15
1.4 DIAGNOSTICS and MODEL EVALUATION c circlecopyrt HYON-JUNG KIM, 2017 1.4 DIAGNOSTICS and MODEL EVALUATION When a simple linear model is selected for an application, one cannot be certain in advance that the model is appropriate to the dataset in question. Note that all the tests, intervals, predictions, etc., are based on believing that the assumptions of the regression hold. To assess whether the assumptions underlying the model seem reasonable, we study the residuals with graphical analyses. That is, we analyze the residuals to see if they support the assumptions of linearity, independence, normality and equal variances. When conducting a residual analysis, we make the following plots; (1) plot of the residuals versus the fitted values.

This preview has intentionally blurred sections. Sign up to view the full version.

This is the end of the preview. Sign up to access the rest of the document.

{[ snackBarMessage ]}

### What students are saying

• As a current student on this bumpy collegiate pathway, I stumbled upon Course Hero, where I can find study resources for nearly all my courses, get online help from tutors 24/7, and even share my old projects, papers, and lecture notes with other students.

Kiran Temple University Fox School of Business ‘17, Course Hero Intern

• I cannot even describe how much Course Hero helped me this summer. It’s truly become something I can always rely on and help me. In the end, I was not only able to survive summer classes, but I was able to thrive thanks to Course Hero.

Dana University of Pennsylvania ‘17, Course Hero Intern

• The ability to access any university’s resources through Course Hero proved invaluable in my case. I was behind on Tulane coursework and actually used UCLA’s materials to help me move forward and get everything together on time.

Jill Tulane University ‘16, Course Hero Intern