Proc logistic model statement default option?

Reply
Occasional Contributor
Posts: 14

Proc logistic model statement default option?

what happens if we exclude the selection option in Proc logistic model statement like below?

proc logistic data=model outmodel=model2;

  class code;

  model possible  (event='1') = Score loss filed amount/lackfit;

  output out = Out2 p = prob xbeta = logit;

run;

thanks,

Super User
Posts: 17,828

Re: Proc logistic model statement default option?

Then it fits the model with all the variables specified.  You've indicated a categorical variable in your class statement but haven't used it.

Occasional Contributor
Posts: 14

Re: Proc logistic model statement default option?

Thanks. So if we have a variable in the Class statement then we need to use that variable in model statement also?

proc logistic data=model outmodel=model2;

  class code;

  model possible  (event='1') = Score Code loss filed amount/lackfit;

  output out = Out2 p = prob xbeta = logit;

run;

Super User
Posts: 17,828

Re: Proc logistic model statement default option?

Yes, if you want it included. For class variables may also want to specify a reference level and your reference type, ie should it be referential or effect coding.

proc logistic data=model outmodel=model2;

  class code/param=ref;

  model possible  (event='1') = Score Code loss filed amount/lackfit;

  output out = Out2 p = prob xbeta = logit;

run;

Here are some quick tips regarding logistic regression in SAS.

Logistic Regression – Quick Tips | Statistics and other stuff from a geek

Occasional Contributor
Posts: 14

Re: Proc logistic model statement default option?

Could you please also help me to understand the use of class statement in Proc Logistic and interpret the table below (table below is from Outmodel=model2) . Thanks.

_TYPE__NAME__CATEGORY__NAMEIDX__CATIDX__MISC_
L ..0
MNYYNYNNN ...
GCodeA002
GCodeB012
GCodeM022
GCodeS032
GCodeU04-2
GCode -1-10
GCode -1-2-1
Gpossible010-11
Gpossible11111
Gpossible -205551258
Gpossible -21662774
Gpossible -2-2-12
ZScore 0.1
Zloss 1.1
Zfiled 2.1
Zamount 3.1
EInterceptE001.13
EEFFECTZ000
EEFFECTE00-0.19
EEFFECTZ101
EEFFECTE10-0.29
EEFFECTZ202
EEFFECTE20-4.87
EEFFECTZ303
EEFFECTE300
EEFFECTV.00
EEFFECTV.10
EEFFECTV.20
EEFFECTV.30
EEFFECTV.40
EEFFECTV.50
EEFFECTV.60
EEFFECTV.70
EEFFECTV.80
EEFFECTV.90
EEFFECTV.100
EEFFECTV.110
EEFFECTV.120
EEFFECTV.130
EEFFECTV.140
X43 281226214093.49
Super User
Posts: 17,828

Re: Proc logistic model statement default option?

I generally don't use that output so I won't comment on it.

Are you unfamiliar with logistic regression or logistic regression in SAS?

If it's in SAS, it's worth reading through the documentation example, specifically the categorical one:

SAS/STAT(R) 9.2 User's Guide, Second Edition

A second good reference is:

Statistical Computing Seminars: Introduction to SAS proc logistic

Occasional Contributor
Posts: 14

Re: Proc logistic model statement default option?

Thanks.

Ask a Question
Discussion stats
  • 6 replies
  • 481 views
  • 0 likes
  • 2 in conversation