cancel
Turn on suggestions
Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.
Showing results for 
Search instead for 
Did you mean: 
giant_wolf00
‎11-08-2017
  • 4 Posts
  • 2 Likes Given
  • 0 Solutions
  • 0 Likes Received

SAS Miner - Impute new scoring data?

by Fluorite | Level 6 in SAS Data Science
‎11-07-2017 04:35 AM
‎11-07-2017 04:35 AM
Hi,   I'm relatively new to SAS Miner so please forgive me if this is a really stupid question!    I created a model which uses 18 variables in it, of which 9 are imputed variables, due to some columns having a high proportion of nulls.  Based on the scoring outputs of the test partition, the created model looked to be fairly predictive (50% of those which had the outcome I was trying to predict, featured in the 10% of scores, c80% in the top 20%).    However, when I came to score some brand new data, whilst the top decile still performed ok (c.50% of the top decile had the outcome I was trying to predict vs. an overall 28%), there were large swathes of records with identical model scores which means some of the "middle" deciles are not performing as expected as they are smeared in the middle.  There are similar levels of nulls in this data too and it is these nulls and the fact that I have used imputed columns in the model creation which prompts my question.   When scoring new data - does Miner factor in the previously used impute, or do I need to feed the new data through an impute before scoring too?   If the former - could something else be wrong which is causing my issue?    Below is my model diagram - the model is the flow on the right, from data all the way to scoring the test partition.  The flow on the left (starting highlighted yellow), is the new data I'm trying to score.   Any help, very gratefully received.  Thank you.   ... View more

Re: Miner - Scoring New Data - Scored variable not found

by Fluorite | Level 6 in SAS Data Science
‎10-13-2017 10:12 AM
‎10-13-2017 10:12 AM
Wendy,   Thank you..... its fixed!  It wouldn't let me do it exactly how you described (not sure why - unless I wrote it wrongly), so I created an additional code node in between using:   data mydata.mytable; Set &em_import_score; run;   I then carried on as normal using the proc sort data code, and it appears to have worked, so I'm very grateful to you.   Thank you. ... View more

Re: Miner - Scoring New Data - Scored variable not found

by Fluorite | Level 6 in SAS Data Science
‎10-13-2017 06:05 AM
‎10-13-2017 06:05 AM
Wendy,   that's great thank you - I really appreciate the reply and help. I have "browsed" the data (I learnt something new!) and there are indeed scored columns in there.  However they aren't in the formt I was expecting for the column names - which is possibly where it is going wrong in the code node when I come to sort.  I want to rank (and decile) the predicted outcome = 'Y', which presumably is the column "predicted: DEBT_90_DAYS_ANY=Y".      My code for the sort is referencing "P_DEBT_90_DAYS_ANYY"  as I thought the format was p_mycolumnnameY (or 1/0 if using 1 or 0 instead of Y/N).  However it says the variable doesn't exist, so I'm guessing I have the naming convention wrong or I've missed a step somewhere.   proc sort data= mydata.my_original_table; out= mydata.my_original_table; by descending P_DEBT_90_DAYS_ANYY; run;   The data table referenced in the sort code is the same table name as the table that was imported with the new data I wanted to score - presumably this is the correct way?   Thanks in advance for any further help,   GW ... View more

Miner - Scoring New Data - Scored variable not found

by Fluorite | Level 6 in SAS Data Science
‎10-12-2017 11:08 AM
‎10-12-2017 11:08 AM
Hi,   Apologies, I'm REALLY new to SAS, so if this is a basic error forgive me, however I can't figure out where I've gone wrong or indeed neither can anyone else in my organisation.    I have created a Logistic Regression stepwise model to predict someone's likelihood to get into debt.  The model has some grouped and imputed columns in the final selection.  It was also partitioned into training, validation, and test Partitions.    I have successfully scored the test partition data without issue, and now I want to score a brand new set of data - it is with this I have the issue.    The new data has been imported, the roles and data levels (nominal, internal...) have all been changed to the appropriate type and match the data set the model was originally built on.  The role of the new data is set to "score" and the subsequent score node appears to have run without issue (it has a green tick!).  I then have a code node attached to this to sort and rank (decile) the scored column, however it says the scored variable P_column_nameY (my output is Y/N) doesn't exist.   Just to double check I didn't get the variable name wrong I then attached a code node to the score node to "print" the data...   "proc print data= tablename;  run;"     The output of the "print", doesn't include any scored columns at all, so no wonder the sort and rank wasn't working.  Does anyone know why and where I've gone wrong?   I have attached a screen grab of my miner project - the bit within the yellow box the process from start to finish.  Anything outside of this are different variations of the model I haven't taken forward or the (working) scoring of the modelled test data     Thanks in advance,   GW.       ... View more
Contact Me
Online Status
Offline
Date Last Visited
‎11-08-2017 07:04 AM
Likes given to
View All
Quick links