Binning (categorize continuous var into categories)

[Ronein]

run your code on my data on specific variable (Nr Years in bank ) 

In your code the best group should be 1 and worse should be 4

I don't see that Per increasing consistently by moving from group 1 to 2 and 2 to 3 and 3 to 4.

Is it not part of the requirement??? To have consistently increasing in per (percentage of bads)???

[Ksharp]

"I don't see that Per increasing consistently by moving from group 1 to 2 and 2 to 3 and 3 to 4."
You should calcualte it on your own, my code did not include it.
bad_pct=n_bad/n;
====>
group bad_pct
1 0.062
2 0.036
3 0.02
4 0.011

[Ronein]

The code you send me is running 100%

It provide good results but my friend provide better results via Miner (IGN)

It would be best if  can provide such results in SAS code.

By the way- what is the alogorithm name that you used in your code?

[Ksharp]

"It provide good results but my friend provide better results via Miner (IGN)"
You could try SAS/EM 's ScoreCard module,but that need money to pay.
And I never heard of Miner ,what kind of software it is?

"It would be best if can provide such results in SAS code."
1)You could ask @RobPratt to write SAS/OR code for it ,and also need to pay money to get SAS/OR.
2)Make %let n_iter=10000 ; as big as you can , but that maybe cost you an entire day to get better result.

And make the following 1000 as big as you can. E.X. 10000

call gainit(id,1000,encoding);
--->
call gainit(id,10000,encoding);

"what is the alogorithm name that you used in your code?"
Genetic Algorithm. Check @Rick_SAS blog:
https://blogs.sas.com/content/iml/2021/10/20/intro-genetic-algorithms-sas.html
https://blogs.sas.com/content/iml/2021/10/18/crossover-mutation.html

[Ksharp]

"What is the meaning of values "M" and "I"?"
It is means negative infinity and positive infinity .
A.K.A you can map your data into bins by these cutpoints:
low-137='1'
137-144='2'
144-174='3'
174-high='4'

"I also added to final table percent of bads within each group"
And also glad to see you make my code running to get the perfect result . Congratulations!

[Ksharp]

And about your time costing problem,
You could start 10 sas session ,each session for ONE variable,that could save you lots of time.
P.S. if you could set "%let n_iter=100" as big as you can to get the better IV.

[Ksharp]

"M" and "I"
represent two missing value .M and .I ,
.M stands for negative infinity,
.I stands for positive infinity.

[Ronein]

I don't know exactly about miner .
I know that it belong to sas.
Sas enterprise miner

[Ronein]

My friend told me that SAS/EM 's ScoreCard module is within sas miner

[RobPratt]

You might be interested in PROC OPTBINNING:

SAS Help Center: OPTBINNING Procedure