Hello, I am trying to prepare an extrapolated modeling dataset, which is quite challenging because there are overlapping periods. I have written a query below, and I am close, but could you please help me reach the attached target table?   DATA HAVE; LENGTH ID 8 APP_DATE 8 RATING $ 10 DELINQUENCY 8 BALANCE 8 ; INFILE DATALINES MISSOVER DLM=","; INPUT ID APP_DATE RATING DELINQUENCY BALANCE; FORMAT APP_DATE DATE9.; DATALINES; 1,20027,BB+,0,10000 1,20119,BB,30,9000 1,20635,BB-,60,8000 1,20939,Default,91,5000 2,19023,BB-,30,7000 2,19509,BB,0,6000 2,20027,BB-,30,4000 2,20147,Default,91,3000 RUN; DATA INTERMEDIATE; SET HAVE; FORMAT DATE_OF_DEFAULT DATE9.; IF RATING = 'Default' THEN DATE_OF_DEFAULT=INTNX("month",APP_DATE,0,"E"); RUN; PROC SORT DATA=INTERMEDIATE; BY ID DESCENDING DATE_OF_DEFAULT APP_DATE; RUN; DATA WANTED; SET INTERMEDIATE; BY ID DESCENDING DATE_OF_DEFAULT APP_DATE; RETAIN DATE_OF_DEFAULT_; FORMAT DATE_OF_DEFAULT_ DEFAULT_W12MNTH DEFAULT_W24MNTH COHORT_DATE DATE9.; IF NOT MISSING(DATE_OF_DEFAULT) THEN DATE_OF_DEFAULT_=DATE_OF_DEFAULT; ELSE DATE_OF_DEFAULT=DATE_OF_DEFAULT_; MONTH_DIFFERENCE=intck("MONTH", APP_DATE, DATE_OF_DEFAULT); DEFAULT_W12MNTH= INTNX("MONTH",DATE_OF_DEFAULT ,-12,'S'); DEFAULT_W24MNTH= INTNX("MONTH",DATE_OF_DEFAULT ,-24,'S'); IF MONTH_DIFFERENCE>0 THEN DO; DO I=0 TO MONTH_DIFFERENCE; COHORT_DATE= INTNX("MONTH",APP_DATE ,I,'E'); IF COHORT_DATE > DEFAULT_W12MNTH THEN DEFAULT_FLG12MNTH=1; ELSE DEFAULT_FLG12MNTH=0; IF COHORT_DATE > DEFAULT_W24MNTH THEN DEFAULT_FLG24=1; ELSE DEFAULT_FLG24=0; OUTPUT; END; END; ELSE DO; COHORT_DATE=DATE_OF_DEFAULT; OUTPUT; END; DROP DATE_OF_DEFAULT_ I MONTH_DIFFERENCE DEFAULT_W12MNTH DEFAULT_W24MNTH; RUN; PROC SORT DATA=WANTED ; BY ID APP_DATE DATE_OF_DEFAULT ; RUN; ... View more

Hi everybody,   I can calculate BINOM.DIST with the PDF function in SAS, but I'm looking for the equivalent of the BINOM.DIST.RANGE function. Let me give you an example;   DATA Have; Length Bucket 8 Obs 8 Defaults 8 Avg_PD 8 ; Infile Datalines Missover Dlm=","; Input Bucket Obs Defaults Avg_PD; Format Avg_PD Percent8.2; Datalines; 1,300,5,0.01 2,1000,20,0.02 3,2000,80,0.04 4,2000,100,0.05 5,2000,150,0.07 6,1000,80,0.08 7,800,100,0.12 8,200,20,0.14 ; DATA WANT; SET HAVE; DR=Defaults/Obs; PValue=PDF('BINOMIAL',Defaults/*number of successes*/,DR /*probability of success*/,Obs/*number of trails*/); Run; I'm trying to get the same results in SAS as in the Excel table below.   Bucket Obs Defaults Avg PD DR P-Value 1 300 5 1,0% 1,7% 18,39% 2 1000 20 2,0% 2,0% 53,06% 3 2000 80 4,0% 4,0% 51,58% 4 2000 100 5,0% 5,0% 51,43% 5 2000 150 7,0% 7,5% 20,14% 6 1000 80 8,0% 8,0% 51,67% 7 800 100 12,0% 12,5% 34,73% 8 200 20 14,0% 10,0% 96,33%   PValue = =BINOM.DIST.RANGE(Obs;AvgPd;Defaults;Obs)   Your help is greatly appreciated; thanks in advance. ... View more

Hello, Since there is no topic about modeling approaches, I am asking my question under this topic. My question will be about LGD modeling. We have some input variables for Loss Rate calculation in LGD modeling. These are; Payment: Monthly regular payments of the loan Indirect cost Direct cost Drawings Can you explain with examples for the parts other than Payment? Also, can you explain how Discounting is calculated? Is there a similar web platform where I can ask such questions?   Thank you ... View more

Hello everbody,   I would like to ask that how can we make anomaly detection - unsupervised learning in SAS 9.4 without VIYA. Is there way to do it? If there is no way to do it without SAS VIYA license, what can be alternative methods? The most important question is “how?”, how can make some beneficial anomaly analysis in SAS 9.4? Can someone show me with some samples? Is it possible?   And I want to prevent possible bias results, how can we understand whether the resutls may be BIAS or not?   Thank you, ... View more

Hello everybody,   I try to figure out how can we create a model without a target variable. Do we have to have a Target variable to develop a model? Is it possbile to develop a model without Target? We can use interval Target but are there specific measures for Interval Targets such as it should include three or four distinct values like 1,2,3,4 or nominal like Red, Yellow, Green? Can we elaborate on the definition of Interval Target here?   Also can we create a fake Target?   How can we analyze the data without having a Target variable to interpret the data for future? for forecasting? What are the most commonly used methods? Can Market Basket Analysis be used in this procces?   How can we use Memory Based Reasoning node in this kind of process? Can we also elaborate on the definition of MBA here? We need to understand the logic of MBR well, do you have a samples about this method?   Thanks in advance 🙂 ... View more

Hello everybody,   I would like to ask and learn some approaches, analyzes about representativeness of modeling data. I want to know what is the optimal date range for model development population?   What kind of difference can be between 5 years data set and 1 years data, which conditions should I consider about it? What kind of effect does PSI have? Can it be use for data representation? What are the issues to be considered for non missing rate?  As far as we know, missing rate is always important to define begining point of data?   http://support.sas.com/resources/papers/proceedings10/288-2010.pdf   I need analysis to support the determined population statistically, can somebody help me and give some detailed information about it, please?   Thanks ... View more

Hello everybody,   I need to do some analysis with Text Miner in SAS Enterprise Miner. I have a Text variable which has a 500 character length and this variable has 30000 distinct values, I want to standartize this variable by excluding +/- signs or conjuctions like “or”, “with” and etc. I also want to transform this values from complicated value to pure value, let’s pretend that I have a value as below; Papers *? PAPER paper pAPER Paper pape PAPE Pape Paper with Paper or book   I want to see the above values only Paper, how I can do it with Text Miner? Can somebody hepl me to resolve this, please Thanks ... View more