10-13-2015 06:08 AM
I would like to know if the Basel II reports attached are available on SAS or not. Sometimes titles are different. I would like to know if there are reports with different names. If these reports don't exist in SAS, can you help me how to do it on SAS?
I have checked the link below.
- Confidence Levels for the Gini Coefficient and AUC
- Spiegelhalter Test
- Binomial Calibration Test
- Chi-square (or Hosmer-Lemeshow) Test
- PSI, CSI
- Conditional Entropy Ratio and Information Value, Mann Whitney - U*
- LGD:, “Expected Loss” Shortfall
- T Test
- Cohen’s d-Test
- Descriptive statistics for input variables
10-15-2015 03:59 AM - edited 10-15-2015 04:04 AM
I don't see any attachments and I may be able to help as I use SAS for Basel II at a bank.
Please provide more information. Are you trying to derive PD and LGD, validate them, or do credit risk calculations with them?
11-03-2015 03:02 AM
Thank you for your interest.For instance how can i create following images on SAS. These are T-test and Cohens Test reports and i would like to ask how i can find Confidence Level(%95.0) on SAS which i can find on Excel by using Descriptive Statistics. I found all formulas which i wanted to find by using Proc Means except Confidence Level(%95.0). I couldn’t equalize it with UCLM and LCLM.
11-05-2015 05:54 PM
As far as i understand to create Ttest we need to use Proc Ttest procedure but i need to generate exactly the same output which i posted above. I tried the code as below.
proc ttest data=Rmsaslib.TtestCombine sides=2 alpha=0.01 h0=0; title "Two sample t-test example"; class Formulas; var G1THK; ods output statistics = outputdf ;
ods output equality=outputeq;
ods output TTESTS = PVAL;
ods output ConfLimits = PVAL2; run;
I investigate deeply the Cohens d Test and i didn't find. I guess SAS does not have any procedure related to Cohens d test report. I calculated the columns by writing SAS Code but as i said it before my real purpose is to create exactly the same output. Are there any easiest way to generate it.
proc sql; Create table CohenMerge as select * ,SQRT(sum(SSDF1,SSDF2)/sum(countBacktest,CountTahminUretim)) as S ,((MeanTahminT-MeanTahminB)/calculated S) as Omega ,ABS(calculated omega) AS AbsOmega ,SQRT(((CountTahminUretim+countBacktest)/(CountTahminUretim*countBacktest))+((calculated Omega)**2/(2*(CountTahminUretim+countBacktest)))) as Stddev ,(calculated omega)-1.96*SQRT(calculated Stddev) as ConfInf95G1THK ,(calculated omega)+1.96*SQRT(calculated Stddev) as ConfInf95G2THK from SumSDFG1Table T1 Left Join SumSDFG2Table T2 on T2.obs=T1.obs; quit;
Lastly, Did somebody understand my Confidence Level(%95.0) ?