This preview shows page 1. Sign up to view the full content.
Unformatted text preview: r linearity, normality, and constancy of
variance. • D escribe the shortcomings of the model and what you might do to
o vercome these s hortcomings. • •
• Based on the plot of standardized residuals vs. DealerCost, there are two to
outliers whose standardized value is higher than +2 (214 and 215) and one whose
value is less than -2 (223). The same is clear from the qqplot.
Based on the plot of standardized residuals vs. DealerCost, we see that there is a
pattern in the data and the assumption of variance constancy is not met.
Based on the plot of residuals vs. fitted observation 223 is a leverage (X value
way above the mean of X) and also a standardized residual (Z values less than -2). P art b)
• M ake a log transformation of the dealer cost and the suggested retail
• C reate an R output for the prediction of the log(suggested retail
p rice) from
l og(dealer cost).
> m2 <- lm(log(SuggestedRetailPrice)~log(DealerCost))
lm(formula = log(SuggestedRetailPrice) ~ log(DealerCost))
-0.062920 -0.008694 0.000624 0.010621 0.048798
Estimate Std. Error t value Pr(>|t|)
(Intercept) -0.069459 0.026459 -2.625 0.00924 **
log(DealerCost) 1.014836 0.002616 387.942 < 2e-16 ***
--Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.01865 on 232 degrees of freedom
Multiple R-squared: 0.9985, Adjusted R-squared: 0.9985
F-statistic: 1.505e+05 on 1 and 232 DF, p-value: < 2.2e-16
• ( Y) and interpret the estimated coefficient of log(dealer cost) Foe every one percent increase in the DealerCost, there will be on 1.015% increase in the
Suggested Retail Price.
• C reate the relevant plots to check for lineari ty, normality, and
c onstancy of v ariance. D escribe the improvements...
View Full Document
This note was uploaded on 02/05/2014 for the course STAT 101A taught by Professor Mahtashesfandiari during the Fall '11 term at UCLA.
- Fall '11