Sample 25005: Compute a diagnostic similar to Cook's D for generalized linear models
 |  |  |  |  |  |
Compute a diagnostic similar to Cook's D for generalized linear models
| Contents: | Purpose / Requirements / Details |
NOTE: Beginning in SAS 9.2, this statistic is available both for independent-observation data and clustered data. See the COOKD= and CLUSTERCOOKD= options in the OUTPUT statement in PROC GENMOD.
- PURPOSE:
- Illustrate computing a diagnostic similar to Cook's D statistic for generalized linear models. This diagnostic gives a measure of the influence of each observation on the model fit.
- REQUIREMENTS:
- Base SAS and SAS/STAT Software, Version 7 or later.
- DETAILS:
- Ideally, the influence of an observation on the model fit would be determined by fitting the model with and without the observation and computing twice the difference in the resulting log likelihoods, divided by the number of parameters in the model. However, this would require the model to be fit as many times as there are observations. The statistic computed here is an approximation that is considerably less expensive to compute.
These sample files and code examples are provided by SAS Institute Inc. "as is" without warranty of any kind, either express or implied, including but not limited to the implied warranties of merchantability and fitness for a particular purpose. Recipients acknowledge and agree that SAS Institute shall not be liable for any damages whatsoever arising out of their use of this material. In addition, SAS Institute will provide no support for the materials contained herein.
See the Results tab for the results of running the following SAS code.
data berkeley;
input dept sex yes no total @@;
prop=yes/total;
cards;
1 1 512 331 843 1 2 89 19 108
2 1 353 207 560 2 2 17 8 25
3 1 120 205 325 3 2 202 391 593
4 1 138 279 417 4 2 131 244 375
5 1 53 138 191 5 2 94 299 393
6 1 22 351 373 6 2 24 317 341
;
proc genmod data=berkeley;
class dept sex;
model yes/total=dept sex / dist=binomial link=logit;
output out=obsout stdxbeta=std hesswgt=hesswgt
stdreschi=stdreschi p=pred;
run;
data cookd;
set obsout;
obs=_N_;
/* p is number of parameters excluding the intercept:
5 for DEPT, 1 for SEX */
p=6;
/* h is the leverage */
h=Std*Hesswgt*Std;
cookd=h*Stdreschi**2/((p+1)*(1-h));
run;
proc plot;
plot cookd*obs;
run;
proc print noobs;
var dept sex prop pred cookd;
run;
/********************************************************************
Note that an approximate Cook's D can also be obtained using PROC
LOGISTIC for binary response models. The C= option on the OUTPUT
statement outputs a statistic that is the Cook's D approximation
times the number of parameters.
********************************************************************/
proc logistic data=berkeley;
class dept sex / param=glm;
model yes/total=dept sex;
output out=lout p=pred c=c;
run;
data lout;
set lout;
/* Divide by the total number of parameters in the model */
cookd=c/7;
run;
proc print noobs;
var dept sex prop pred cookd;
run;
These sample files and code examples are provided by SAS Institute Inc. "as is" without warranty of any kind, either express or implied, including but not limited to the implied warranties of merchantability and fitness for a particular purpose. Recipients acknowledge and agree that SAS Institute shall not be liable for any damages whatsoever arising out of their use of this material. In addition, SAS Institute will provide no support for the materials contained herein.
- RESULTS:
-
Following is the output from the PRINT procedure below giving the
values of the diagnostic under the variable COOKD:
DEPT SEX PROP PRED COOKD 1 1 0.60735 0.62914 24.6544 1 2 0.82407 0.65402 0.8009 2 1 0.63036 0.63142 0.2322 2 2 0.68000 0.65622 0.0008 3 1 0.36923 0.33493 0.6589 3 2 0.34064 0.35944 1.7236 4 1 0.33094 0.32815 0.0117 4 2 0.34933 0.35243 0.0105 5 1 0.27749 0.23813 0.2929 5 2 0.23919 0.25831 1.1085 6 1 0.05898 0.06131 0.0123 6 2 0.07038 0.06784 0.0124
Illustrate computing a diagnostic similar to Cook's D statistic for
generalized linear models. This diagnostic gives a measure of
the influence of each observation on the model fit.
| Type: | Sample |
| Topic: | SAS Reference ==> Procedures ==> GENMOD SAS Reference ==> Procedures ==> LOGISTIC |
| Date Modified: | 2015-09-18 09:57:15 |
| Date Created: | 2005-01-13 15:03:22 |
Operating System and Release Information
| Product Family | Product | Host | SAS Release | |
| Starting | Ending | |||
| SAS System | SAS/STAT | z/OS | ||
| Z64 | ||||
| OpenVMS VAX | ||||
| Microsoft® Windows® for 64-Bit Itanium-based Systems | ||||
| Microsoft Windows Server 2003 Datacenter 64-bit Edition | ||||
| Microsoft Windows Server 2003 Enterprise 64-bit Edition | ||||
| Microsoft Windows XP 64-bit Edition | ||||
| Microsoft® Windows® for x64 | ||||
| OS/2 | ||||
| Microsoft Windows 8 Enterprise 32-bit | ||||
| Microsoft Windows 8 Enterprise x64 | ||||
| Microsoft Windows 8 Pro 32-bit | ||||
| Microsoft Windows 8 Pro x64 | ||||
| Microsoft Windows 8.1 Enterprise 32-bit | ||||
| Microsoft Windows 8.1 Enterprise x64 | ||||
| Microsoft Windows 8.1 Pro | ||||
| Microsoft Windows 8.1 Pro 32-bit | ||||
| Microsoft Windows 95/98 | ||||
| Microsoft Windows 2000 Advanced Server | ||||
| Microsoft Windows 2000 Datacenter Server | ||||
| Microsoft Windows 2000 Server | ||||
| Microsoft Windows 2000 Professional | ||||
| Microsoft Windows NT Workstation | ||||
| Microsoft Windows Server 2003 Datacenter Edition | ||||
| Microsoft Windows Server 2003 Enterprise Edition | ||||
| Microsoft Windows Server 2003 Standard Edition | ||||
| Microsoft Windows Server 2003 for x64 | ||||
| Microsoft Windows Server 2008 | ||||
| Microsoft Windows Server 2008 R2 | ||||
| Microsoft Windows Server 2008 for x64 | ||||
| Microsoft Windows Server 2012 Datacenter | ||||
| Microsoft Windows Server 2012 R2 Datacenter | ||||
| Microsoft Windows Server 2012 R2 Std | ||||
| Microsoft Windows Server 2012 Std | ||||
| Microsoft Windows XP Professional | ||||
| Windows 7 Enterprise 32 bit | ||||
| Windows 7 Enterprise x64 | ||||
| Windows 7 Home Premium 32 bit | ||||
| Windows 7 Home Premium x64 | ||||
| Windows 7 Professional 32 bit | ||||
| Windows 7 Professional x64 | ||||
| Windows 7 Ultimate 32 bit | ||||
| Windows 7 Ultimate x64 | ||||
| Windows Millennium Edition (Me) | ||||
| Windows Vista | ||||
| Windows Vista for x64 | ||||
| 64-bit Enabled AIX | ||||
| 64-bit Enabled HP-UX | ||||
| 64-bit Enabled Solaris | ||||
| ABI+ for Intel Architecture | ||||
| AIX | ||||
| HP-UX | ||||
| HP-UX IPF | ||||
| IRIX | ||||
| Linux | ||||
| Linux for x64 | ||||
| Linux on Itanium | ||||
| OpenVMS Alpha | ||||
| OpenVMS on HP Integrity | ||||
| Solaris | ||||
| Solaris for x64 | ||||
| Tru64 UNIX | ||||