Programming the statistical procedures from SAS

Proc Logistic offset= value not saving in outmodel= coefficients file

Reply
Frequent Contributor
Posts: 101

Proc Logistic offset= value not saving in outmodel= coefficients file

 

 

 

I’m using an offset variable in some model builds and I am trying to output the coefficients in order to score datasets for automated scoring.  When I go to score the data it seems like the offset value is not saving in the coefficients file. 

 

In the past, when I have used the WEIGHT option, I never had this problem. 

 

Is there I when to save the offset value without having to calculate the  “off”  variable everytime I score data?

 

proc logistic data= training outmodel= mod.MYMODEL descending;
model target=  
sales_ratiod
AGE_GT45
EM
RECENCY
PURCHASE_ORDERS
X_division
INDL_BT
WEEKEND_SHOPPER
/ offset=off   ; 
output predicted=p_response; run; quit;
 
 
 
proc logistic inmodel=MYMODEL   ;
score data= MYDATA  out=SCORED_DATA;
run;
 
ERROR: Variable off not found in DATA=MYMODEL  
NOTE: The SAS System stopped processing this step because of errors.
WARNING: The data set WORK.SCORED_DATA may be incomplete.  When this step was stopped there were 0 observations and 0 variables.
WARNING: Data set WORK.SCORED_DATA was not replaced because this step was stopped.
NOTE: PROCEDURE LOGISTIC used (Total process time):
      real time           0.00 seconds
      cpu time            0.01 seconds

 

Super User
Posts: 18,591

Re: Proc Logistic offset= value not saving in outmodel= coefficients file

OFFSET may be a fixed value for you, but in the program it's designed to accept a variable. So it's not being treated as a constant, which is why it needs to be specified in the scoring data as well. It could be different values dependent on input variables for example.

Super User
Posts: 9,779

Re: Proc Logistic offset= value not saving in outmodel= coefficients file

Could you try SCORE statement in PROC LOGISTIC.

 

proc logistic data= training outmodel= mod.MYMODEL descending;
model target=  
sales_ratiod
AGE_GT45
EM
RECENCY
PURCHASE_ORDERS
X_division
INDL_BT
WEEKEND_SHOPPER
/ offset=off   ; 

score data= MYDATA  out=SCORED_DATA;
run;
Frequent Contributor
Posts: 101

Re: Proc Logistic offset= value not saving in outmodel= coefficients file

Thanks Ksharp, 

 

Unfortunately, the off var still did not  save in the coefficients file. 

 

 

23         proc logistic data=w.training_adj outmodel= mod.MYMODEL NAMELEN =32 descending;

24         model winback=

25         sales_ratiod

26         AGE_GT45

27         EM

28         RECENCY

29         PURCHASE_ORDERS

30         X_division

31         INDL_BT

32         WEEKEND_SHOPPER

33         DISTANCE_IND

34         PURCHASES_Q1

35         INDP_FOOTWEAR_B

36         INDP_SWEATERS_M

37         INDP_FOOTWEAR_M

38         DOLL_HOME

39         INDP_OUTERWEAR_B

40         West_ind

41         AUR

42         FTRANS

43         INDL_GOLF_TENNIS_W

44         FINDL_ELEVATED_W

45         INDP_DRESSES_SKIRTS_W

46         / stb lackfit hier=single  offset=off  /*ctable pprob=0.1 to .40 by .01*/ ;

47         output out=out predicted=p_response;

48         score data= VALIDATION_ADJ  out=SCORED_DATA; run;

 

ERROR: Variable off not found in DATA=WORK.VALIDATION_ADJ.

NOTE: The SAS System stopped processing this step because of errors.

NOTE: There were 1444748 observations read from the data set W.TRAINING_ADJ.

WARNING: The data set MOD.MYMODEL may be incomplete.  When this step was stopped there were 0 observations and 6 variables.

WARNING: Data set MOD.MYMODEL was not replaced because this step was stopped.

 

WARNING: Data set WORK.OUT was not replaced because this step was stopped.
WARNING: The data set WORK.SCORED_DATA may be incomplete. When this step was stopped there were 0 observations and 0 variables.
WARNING: Data set WORK.SCORED_DATA was not replaced because this step was stopped.

SAS Employee
Posts: 245

Re: Proc Logistic offset= value not saving in outmodel= coefficients file

The offset variable is essentially a predictor variable in the model.  Any data that you want to score has to have values of the model predictors, so the offset has to be available.

Ask a Question
Discussion stats
  • 4 replies
  • 204 views
  • 2 likes
  • 4 in conversation