Ch. 8 Nonlinear Regression Functions

Ch. 8 Nonlinear Regression Functions - NONLINEAR REGRESSION...

Info iconThis preview shows pages 1–12. Sign up to view the full content.

View Full Document Right Arrow Icon
1 | Nonlinear NONLINEAR REGRESSION FUNCTIONS (Ch. 8) The recommended exercise questions from the textbook: Chapter 8: All except (8.9), for the 2 nd ed; All except (8.9), (8.11) – (8.12), for the 3 rd ed.
Background image of page 1

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

View Full DocumentRight Arrow Icon
2 | Nonlinear (1) Motivation
Background image of page 2
3 | Nonlinear [Example] Return to our TestScore example. Wish to look at the relation between students’ performances and district income: T e s t S c o r e . AVGINC (district average income, $1,000).
Background image of page 3

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

View Full DocumentRight Arrow Icon
4 | Nonlinear 600 620 640 660 680 700 720 0 10 20 30 40 50 60 AVGINC TESTSCR Set up this model: 2 01 2 TestScore AVGINC AVGINC u   .
Background image of page 4
5 | Nonlinear [Quadratic regression model] Dependent Variable: TESTSCR Method: Least Squares Included observations: 420 White Heteroskedasticity-Consistent SE & Covariance Variable Coefficient Std. Error t-Statistic Prob. C 607.30 2.902 209.3 0.0000 AVGINC 3.851 0.268 14.36 0.0000 AVGINC^2 -0.042 0.005 -8.851 0.0000 R-squared 0.556 Mean dependent var 654.2 Adjusted R-squared 0.554 S.D. dependent var 19.05 Log likelihood -1662.7 F-statistic 261.28 Durbin-Watson stat 0.951 Prob(F-statistic) 0.0000
Background image of page 5

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

View Full DocumentRight Arrow Icon
6 | Nonlinear 600 620 640 660 680 700 720 740 0 10 20 30 40 50 60 AVGINC TESTSCR TESTSCRF_1 TESTSCRF_2 How can we test H o : The correct model is a linear regression model?
Background image of page 6
7 | Nonlinear (2) General model 1) General model: Y = f ( X 1 , . .. , X k ) + u . w h e r e f ( ) is called “nonlinear regression function.” 2) How to estimate the effect on Y of a change in X . • Expected change in Y ( Y ) associated with the change in X 1 ( 1 X ) holding X 2 ,..., X k constant: 11 2 1 2 ( , ,..., ) ( , ,..., ) kk Y f XX f X  . • Estimated change: 1 ˆˆ ( ,..., ) ( ,..., ) Y f XXX f .
Background image of page 7

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

View Full DocumentRight Arrow Icon
8 | Nonlinear [Example] Wish to estimate the change in test scores associated with a change in district income of $1,000, based on the estimated quadratic regression function. 22 01 2 2 12 ˆˆ ˆ ˆ (1 1 1 1 ) 0 1 0 ) 21 . Testscr    How can we estimate () SE Testscr ? • Using eviews, test H 0 : β 1 + 21 β 2 = 0. 21 2.963 Testscr . 1 1 2 2 ˆ ˆ ˆ ˆ var( 21 ) var( ) 42cov( , ) 441var( )  . ( ) 0.17 SE Testscr .
Background image of page 8
9 | Nonlinear Wald Test: Equation: Untitled Test Statistic Value df Prob F-statistic 299.94 (1, 417) 0.000 Chi-square 299.94 1 0.000 Null Hypothesis Summary: Normalized Restric. (=0) Value Std. Err. C(2) + 21*C(3) 2.963 0.171
Background image of page 9

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

View Full DocumentRight Arrow Icon
10 | Nonlinear (3) Nonlinear functions A. Polynomials: Y = β 0 + β 1 X + β 2 X 2 + . .. + β r X r + u . 1) Testing linearity: T e s t H o : β 2 = . .. = β r = 0. 2) Determining maximum r . (General to specific) • Pick a maximum value of r and estimate the model for that r . • Check the significance of β r . If significant, then keep using that r . • If β r is not significant, estimate the model with maximum ( r -1). If β r -1 is significant, then keep using ( r -1). • If β r -1 is not significant, try ( r -2). Repeat this procedure until the last β becomes significant.
Background image of page 10
11 | Nonlinear [Example] (Specific to general approach) 23 01 2 3 TestScore AVGINC AVGINC AVGINC u   .
Background image of page 11

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

View Full DocumentRight Arrow Icon
Image of page 12
This is the end of the preview. Sign up to access the rest of the document.

Page1 / 38

Ch. 8 Nonlinear Regression Functions - NONLINEAR REGRESSION...

This preview shows document pages 1 - 12. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online