topic Re: resid vs fitted plot in SAS Enterprise Guide

resid vs fitted plot

smorkoc — Mon, 10 Sep 2018 00:01:30 GMT

                                      Plot of resid*Flux1.  Legend: A = 1 obs, B = 2 obs, etc.

       resid |
             |
         0.4 +
             |
             |
             |
             |                                                                                                  A
             |
         0.3 +
             |
             |
             |
             |
             |                                                                                                         A
         0.2 +                                                                                A
             |
             |                                                                                                      A
             |                                                                     A
             |                                                                                   A
             |                                 A
         0.1 +              A                                                                 A
             |                  A                   A      A   A         A              A          A         A
             |      A                                       BA
             |       BA   A      A                 A   A       B       A
             |      A  A     A BA    AAB   A  A                    A
             |       AAAA    A A       B  AA           AA          A              A
         0.0 +      AC   B A             AAA                                         A           A
             |       AA A AA   A              A                                            A
             |       DA     A B A                  A A    B
             |           A     A   B  A A A     A   AA  A                                   A
             |                       A
             |         A A                  A    A                  A    A
        -0.1 +                                                    A                 A
             |                                                     A
             |        A                                       AA                 A
             |
             |                                                              A
             |
        -0.2 +                       A                    A      A
             |
             |
             |
             |                                                                 A
             |
        -0.3 +
             |
             ---+--------------+--------------+--------------+--------------+--------------+--------------+--------------+--
              0.00           0.20           0.40           0.60           0.80           1.00           1.20           1.40

                                                                  Flux1

here is my resid vs fitted plot, i was not sure it seems ok, could you help me? what should do next, if I face this type of plot.

Re: resid vs fitted plot

Reeza — Mon, 10 Sep 2018 00:33:21 GMT

What you're looking for:

The residuals "bounce randomly" around the 0 line. This suggests that the assumption that the relationship is linear is reasonable.
The residuals roughly form a "horizontal band" around the 0 line. This suggests that the variances of the error terms are equal.
No one residual "stands out" from the basic random pattern of residuals. This suggests that there are no outliers.

Source: https://onlinecourses.science.psu.edu/stat501/node/277/

What you have:

Residuals that are not bouncing randomly. They are lower around the lower values of flux and more spread out towards the higher end.
No horizontal band, it's more of a funnel shape.
There are several residuals that seem problematic, one high up and one low down.

Next steps:

Attempt to refit the model to see if another variable reduces the error.
Transformation on variables to reduce variance
Something else, that makes sense given the contact of your problem which we only know a few things about.

@smorkoc wrote:

                                      Plot of resid*Flux1.  Legend: A = 1 obs, B = 2 obs, etc.

       resid |
             |
         0.4 +
             |
             |
             |
             |                                                                                                  A
             |
         0.3 +
             |
             |
             |
             |
             |                                                                                                         A
         0.2 +                                                                                A
             |
             |                                                                                                      A
             |                                                                     A
             |                                                                                   A
             |                                 A
         0.1 +              A                                                                 A
             |                  A                   A      A   A         A              A          A         A
             |      A                                       BA
             |       BA   A      A                 A   A       B       A
             |      A  A     A BA    AAB   A  A                    A
             |       AAAA    A A       B  AA           AA          A              A
         0.0 +      AC   B A             AAA                                         A           A
             |       AA A AA   A              A                                            A
             |       DA     A B A                  A A    B
             |           A     A   B  A A A     A   AA  A                                   A
             |                       A
             |         A A                  A    A                  A    A
        -0.1 +                                                    A                 A
             |                                                     A
             |        A                                       AA                 A
             |
             |                                                              A
             |
        -0.2 +                       A                    A      A
             |
             |
             |
             |                                                                 A
             |
        -0.3 +
             |
             ---+--------------+--------------+--------------+--------------+--------------+--------------+--------------+--
              0.00           0.20           0.40           0.60           0.80           1.00           1.20           1.40

                                                                  Flux1

here is my resid vs fitted plot, i was not sure it seems ok, could you help me? what should do next, if I face this type of plot.

Re: resid vs fitted plot

PaigeMiller — Mon, 10 Sep 2018 01:00:37 GMT

Is flux1 truly the fitted value (which can also be called the predicted value), or is it actually the X value used in the regression?

Re: resid vs fitted plot

smorkoc — Mon, 10 Sep 2018 01:05:01 GMT

It is fitted value (predicted value).

Re: resid vs fitted plot

smorkoc — Mon, 10 Sep 2018 01:06:15 GMT

Thanks for the help. I will consider transformations for the variables.