HELP SAS ANOVA output

From: Nick Funk (nfunk_at_rtconline-NOSPAM.com)
Date: 03/23/05 

Date: Wed, 23 Mar 2005 11:57:36 -0600

I am new to SAS and it has been over 30 years since I have done any
statistics. I am having a problem understanding the output from SAS.
Data below is varying suspended sediment concentration and measured
reflectance from 4 different spectrum bands.

I need to select the best band for monitoring the suspended sediment
concentration. From my understanding of the F-statistic, (ratio of
model mean square to error mean square) the band with the lower value of
F-statistic would be the better match. Therfore band 4 would be the
best spectrum band for monitoring sediment concentrations. Is my
premise true?

Any information would be greatly appreciated!
Nick (a very old student).

data ssc;
title 'Landsat Thematic Mapper Bands Simulation';
Input ssc band1 band2 band3 band4;
cards;
0 1.787 1.760 1.351 0.206
50 3.668 4.588 3.837 0.733
100 4.716 6.869 6.236 1.215
150 4.896 7.852 7.792 1.677
200 5.106 8.542 8.939 2.061
250 5.278 8.960 9.535 2.196
300 5.405 9.332 10.265 2.584
350 5.468 9.523 10.803 2.907
400 5.668 9.910 11.419 3.263
450 5.555 9.850 11.601 3.546
500 5.807 10.228 12.049 3.736
550 5.876 10.408 12.323 3.738
600 5.957 10.583 12.889 4.455
650 5.750 10.265 12.555 4.434
700 5.867 10.436 12.868 4.658
750 5.831 10.500 13.111 4.963
800 5.927 10.583 13.221 5.086
850 6.019 10.713 13.438 5.213
900 5.764 10.190 12.880 5.148
950 5.778 10.245 13.026 5.433
1000 5.786 10.280 13.141 5.518
;
run;
proc print;
run;
PROC GLM;
model SSC = band1 band2 band3 band4;
run;

____________________________________________________________________________

          Landsat Thematic Mapper Bands Simulation 26
                                                    11:00 Tuesday, March
23, 2005
 

                                The GLM Procedure

 

Dependent Variable: ssc

 

                                        Sum of

  Source DF Squares Mean Square F Value Pr > F
 
         Model 4 1905594.102 476398.525 392.79 <.0001

Error 16 19405.898 1212.869
 

Corrected Total 20 1925000.000
 

 

                R-Square Coeff Var Root MSE ssc Mean

 

                0.989919 6.965253 34.82626 500.0000

 

 

  Source DF Type I SS Mean Square F Value Pr > F
 

  band1 1 975189.9091 975189.9091 804.04 <.0001
  band2 1 244490.5692 244490.5692 201.58 <.0001
  band3 1 645864.4761 645864.4761 532.51 <.0001
  band4 1 40049.1473 40049.1473 33.02 <.0001
 

 

  Source DF Type III SS Mean Square F Value Pr > F
 

  band1 1 337.58127 337.58127 0.28 0.6050
  band2 1 817.34546 817.34546 0.67 0.4238
  band3 1 2667.62651 2667.62651 2.20 0.1575
  band4 1 40049.14729 40049.14729 33.02 <.0001
 

 

                                          Standard

        Parameter Estimate Error t Value Pr >
|t|
 

        Intercept 24.3775158 112.6279185 0.22
0.8314
        band1 -72.8225245 138.0331632 -0.53
0.6050
        band2 119.5391038 145.6176619 0.82
0.4238
        band3 -126.4894497 85.2901102 -1.48
0.1575
        band4 322.5818530 56.1371267 5.75
<.0001

Relevant Pages

  • Re: OH Bother!!!
    ... progress...called triangles. ... The bottom row has the blocks I ... Could you change the border somehow so it ends up square? ... one band black, one band of scrappy brights, one band black ... ...
    (rec.crafts.textiles.quilting)
  • Re: HELP SAS ANOVA output
    ... Yes, I only want to select the one best spectral band, which I assume is ... >> I am new to SAS and it has been over 30 years since I have done any ... > see which one gives the best fit? ...
    (sci.stat.math)
  • Re: OH Bother!!!
    ... one band black, one band of scrappy brights, one band black ... ... The top isn't square, it's a rectangle that is wider than tall. ... I know I could add an extra row at the bottom by opening up the seam lines a little bit, but honestly, I don't know if I could face sewing another hundred or so triangles together!!! ... I still have to put an applique border on it. ...
    (rec.crafts.textiles.quilting)
  • Orintability of the Möbius Band
    ... The Band becomes quite orientable when the ... thickness of the strip. ... Mobius band connected boundary/edgewise to the main band. ... twisted square section can be looked upon as several thinner Möbius ...
    (sci.math)
  • Re: Band Turmoil
    ... <<It's EXPONENTIALLY harder to start from square one as a new band ... sound, even if most of the music and the players are the same. ... which is why we got the service mark for our band name when we ...
    (alt.guitar.bass)