This was a homework question I did in R. Now, I want to practice SAS coding. So, I want to do it again in SAS to see if output match with what I did in R. I do not know how to fit test data after I fit training data. Here is what I did, SAS 9.4 code I have so far: proc import out=work.car5 datafile= "C:\newRay\uofacourses\Stat441\assignment1\Auto1.csv" dbms=csv; getnames=yes; datarow=2; run; data car6; set work.car5; if mpg>=23 then GC="Y"; else GC="N"; run; data train test; set car6; b=mod(year,2); if b=0 then output train; else output test; run; proc logistic data=train; model GC = cylinders displacement horsepower weight acceleration year; run; So, I import the auto data. I created a new variable called "GC" which means good car with mpg above median. So, GC=Y means yes it is a good car. otherwise is GC=N with mpg below median. I split the data into training and test data based on model year. Even year goes to training data. Odd year goes to test data. Then I fit logistic regression with training data. This code works fine. I do not know what to do next. So, next I want to use the model fitted in training data for test data. Then find predicted outcome and compare to real outcome in test data. For test data, Test error rate=("number of cases predicted as Y but real data is N"+"number of cases predicted as N but real data is Y")/total cases How can I do this whole task in SAS? I have also attach my data file. If anyone can help by providing some useful code or guide, I will appreciate it. Thanks in advance!
... View more