PPT9 Residual Analysis 1 F2010

PPT9 Residual Analysis 1 F2010 - McGill University Advanced...

Info iconThis preview shows pages 1–22. Sign up to view the full content.

View Full Document Right Arrow Icon
McGill University Advanced Business Statistics MGSC-272
Background image of page 1

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

View Full DocumentRight Arrow Icon
Residual Analysis Using residuals to detect departure from assumptions
Background image of page 2
Residual Plots Plots of the residuals in a regression analysis can be used to detect departures from the assumption of: 1.Normality of the error terms 2.Constant variance 3.Independence of the error terms
Background image of page 3

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

View Full DocumentRight Arrow Icon
Definition of Residual ( 29 0 1 1 Error term in a true multiple regression model ( ) ... i i i k k y E y y x x ε β = - = - + + + ( 29 0 1 1 Residual term in an estimated mutiple regression model ˆ ... i i i i k k e y y y b b x b x = - = - + + +
Background image of page 4
Example 8.1 Page 251 ATHLETE FAT INTAKE CHOLESTEROL 1 1290 1182 2 1350 1172 3 1470 1264 4 1600 1493 5 1710 1571 6 1840 1711 7 1980 1804 8 2230 1840 9 2400 1956 10 2930 1954
Background image of page 5

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

View Full DocumentRight Arrow Icon
Calculating Residuals The regression equation is CHOLESTEROL = 578.9 + 0.5403 FAT INTAKE X 1 = 1290 = 578.9 + 0.5403(1290) = 1275.92 1 ˆ y 1 1 ˆ - 1182 -1275.92 -93.92 i e y y = = = Residual
Background image of page 6
Residuals: Linear Model Fitted Value Residual 2200 2000 1800 1600 1400 1200 200 100 0 -100 -200 Residuals Versus the Fitted Values (response is CHOLESTEROL) e 1 = -93.92
Background image of page 7

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

View Full DocumentRight Arrow Icon
??Pattern?? Do you see a pattern in the residual plot? Residuals are positive for athletes with intermediate levels of fat intake. Residuals are negative for athletes with low or high levels of fat intake. This suggests the use of a second order (Quadratic) model.
Background image of page 8
Quadratic Model The regression equation is: CHOLESTEROL = - 1216.14 + 2.3989 FAT INTAKE - 0.000450 FAT^2 X 1 = 1290 2 ˆ 1216.14 2.3989 .00045 y x x = - + - 2 1 ˆ 1216.14 2.3989(1290) .00045(1290) 1129.56 y = - + - = 1 1 1 ˆ 1182 1129.56 52.44 e y y = - = - = Residual
Background image of page 9

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

View Full DocumentRight Arrow Icon
Residuals: Quadratic Model Fitted Value Residual 2000 1900 1800 1700 1600 1500 1400 1300 1200 1100 50 25 0 -25 -50 -75 Residuals Versus the Fitted Values (response is CHOLESTEROL) e 1 = 52.44
Background image of page 10
Detecting Lack of Fit 1. Plot residuals e i on the vertical axis, against each of the independent variables x 1 , x 2 , … x k on the horizontal axis. 2. Plot the residuals on the vertical axis against the predicted value, , on the horizontal axis. 3. In each plot look for trends, dramatic changes in variability, and/or more than 5% of residuals that lie outside 2 std. dev. of 0. Any of these patterns indicates a problem with model fit. ˆ y
Background image of page 11

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

View Full DocumentRight Arrow Icon
Lack of Fit Tests Minitab Help Screen
Background image of page 12
Regression: Chol (Y) on Fat (X)
Background image of page 13

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

View Full DocumentRight Arrow Icon
Options: Lack of Fit Tests – Data Subsetting
Background image of page 14
Regression Analysis: Chol (y) versus Fat (x) The regression equation is Chol (y) = 579 + 0.540 Fat (x) Predictor Coef SE Coef T P Constant 578.9 167.0 3.47 0.008 Fat (x) 0.54030 0.08593 6.29 0.000 S = 133.438 R-Sq = 83.2% R-Sq(adj) = 81.1% Lack of fit test Possible curvature in variable Fat (x) (P-Value = 0.002 ) Possible lack of fit at outer X-values (P-Value = 0.039) Overall lack of fit test is significant at P = 0.002
Background image of page 15

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

View Full DocumentRight Arrow Icon
Fitted Line Plot
Background image of page 16
Fitted Line Plot: Linear plot of Chol (Y) vs Fat (X)
Background image of page 17

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

View Full DocumentRight Arrow Icon
Linear Plot: R 2 = 83.2%
Background image of page 18
Fitted Line Plot – Quadratic Model
Background image of page 19

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

View Full DocumentRight Arrow Icon
Quadratic Plot: R 2 = 97.7%
Background image of page 20
The Quadratic Model P-value < .001 Impressive!
Background image of page 21

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

View Full DocumentRight Arrow Icon
Image of page 22
This is the end of the preview. Sign up to access the rest of the document.

Page1 / 61

PPT9 Residual Analysis 1 F2010 - McGill University Advanced...

This preview shows document pages 1 - 22. Sign up to view the full document.

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