[R] Problems with Durbin Watson and Partial Residual Plots
John Fox
jfox at mcmaster.ca
Wed Nov 3 00:02:52 CET 2004
Dear Cal,
Both functions work fine for me (Windows XP Pro, R 2.0.0 patched, car
1.0-13, though the versions probably aren't significant here):
> mod <- lm(Sales ~ Time, data=Data)
> library(car)
> durbin.watson(mod)
lag Autocorrelation D-W Statistic p-value
1 0.3232896 1.249564 0.01
Alternative hypothesis: rho != 0
> cr.plot(mod, "Time") # makes the plot!
BTW, since there is only one predictor in the model (i.e., Time), the
component+residual plot is pointless -- it's essentially the scatterplot of
Sales vs. Time.
I'm not sure what the problem is. Perhaps you should just try again (or
provide some more information about what you did).
I hope this helps,
John
--------------------------------
John Fox
Department of Sociology
McMaster University
Hamilton, Ontario
Canada L8S 4M4
905-525-9140x23604
http://socserv.mcmaster.ca/jfox
--------------------------------
> -----Original Message-----
> From: Cal Tate [mailto:mathdoc2be at yahoo.com]
> Sent: Tuesday, November 02, 2004 4:46 PM
> To: John Fox
> Subject: RE: [R] Problems with Durbin Watson and Partial
> Residual Plots
>
> John,
>
> Thanks for your reply. The data set is attached.
>
> Cal
>
> John Fox <jfox at mcmaster.ca> wrote:
>
> Dear Cal,
>
> The functions that you mention are in the car package.
>
> The problems that you've encountered seem very odd to
> me. For example, if
> you take a look at durbin.watson.lm(), you'll see that
> the code producing
> the errors is quite straight-forward; it just extracts
> residuals from the
> model and checks for NAs:
>
> residuals <- residuals(model)
> if (any(is.na(residuals)))
> stop("residuals include missing values")
>
> Likewise, what's going on inside of cr.plot.lm() is
> also pretty simple.
>
> Can you send the data set on which this regression was based?
>
> John
>
> --------------------------------
> John Fox
> Department of Sociology
> McMaster University
> Hamilton, Ontario
> Canada L8S 4M4
> 905-525-9140x23604
> http://socserv.mcmaster.ca/jfox
> --------------------------------
>
> > -----Original Message-----> From:
> r-help-bounces at stat.math.ethz.ch
> > [mailto:r-help-bounces at stat.math.ethz.ch] On Behalf
> Of Cal Tate
> > Sent: Tuesday, November 02, 2004 2:24 PM
> > To: r-help at stat.math.ethz.ch
> > Subject: [R] Problems with Durbin Watson and Partial
> Residual Plots
> >
> > I am trying to evaluate a model by using the commands
> > durbin.watson and cr.plot.
> > However, I keep getting errors that I can't figure out. A
> > description follows. Does anyone have a hint as to
> what may be wrong?
> >
> >
> > 1)The Durbin Watson Test. In running the command I kept
> > getting the message "residuals include missing values" when
> > actually this was NOT the case.
> >
> > Example:
> > >durbin.watson(hw8LM0)
> > Error in durbin.watson.lm(hw8LM0) : residuals include
> missing values
> >
> > (2)Partial Residual Plots: Here I kept getting that the
> > variables I choose are not in the model, i.e. "Time!
> is not in
> > the model" when it clearly is.
> >
> > Example:
> > >cr.plot(hw8LM0,variable="Time")
> > Error in cr.plot.lm(hw8LM0, variable = "Time") :
> > Time is not in the model.
> >
> > Here is the model and residuals:
> > summary(hw8LM0)
> > Call:
> > lm(formula = Sales ~ Time)
> > Residuals:
> > Min 1Q Median 3Q Max
> > -516.32 -292.95 -29.15 238.48 895.09
> > Coefficients:
> > Estimate Std. Error t value Pr(>|t|)
> > (Intercept) 1418.872 122.464 11.59 2.35e-14 ***
> > Time 73.278 4.962 14.77 < 2e-16 ***
> > ---
> > Signif. codes: 0 `***' 0.001 `**' 0.01 `*' 0.05 `.' 0.1 ` '
> > 1 Residual standard error: 389.8 on 40 degrees of freedom
> > Multiple R-Squared: 0.845, Adjusted R-squared: 0.8411
> > F-statistic: 218.1 on 1 and 40 DF, p-value: < 2.2e-16
> >
> > hw8LM0$res
> > 1 2 3 4 5
> > 6
> > 242.679535 679.531369 895.093204 442.975038 -237.4! 43127
> > 245.868708
> > 7 8 9 10 11
> > 12
> > 82.540542 -13.347623 -209.325789 161.976046 -96.612120
> > -271.380285
> > 13 14 15 16 17
> > 18
> > -460.888451 -113.606616 -311.494782 -417.352947 -516.321113
> > 1.430722
> > 19 20 21 22 23
> > 24
> > -18.407443 -328.425609 -476.743774 8.528060 67.849895
> > -298.548271
> > 25 26 27 28 29
> > 30
> > -478.826436 225.895398 110.617233 -226.800933 -487.939098
> > 281.782736
> > 31 32 33 34 35
> > 36
> > -61.495429 -390.773595 -485.051760 431.670075 477.391909
> > -39.886256
> > 37 38 39 40 41
> > 42
> > -276.164422 732.557413 618.279247 -16.998918 -229.277084
> > 756.444751
> >
> >
> >
> > ---------------------------------
> >
> > [[alternative HTML version deleted]]
> >
> > ______________________________________________
> > R-help at stat.math.ethz.ch mailing list
> > https://stat.ethz.ch/mailman/listinfo/r-help
> > PLEASE do read the posting guide!
> > http://www.R-project.org/posting-guide.html
>
>
>
> ________________________________
>
More information about the R-help
mailing list