SUPPORT / SAMPLES & SAS NOTES
 

Support

Problem Note 64302: A scoring test in SAS® Model Manager uses the default-language model instead of a customer-sentiment model

DetailsHotfixAboutRate It

You can use SAS Model Manager to create a test for a project that contains a customer-sentiment model. However, the code that is generated is incorrect. The code that is generated is for a default-language model, whereas the generated code should be for a customer-sentiment model. The generated code should use model={}, instead of language={}.

There are no errors or warnings to indicate that the generated code is incorrect.

To work around the problem, manually modify the code and specify a customer-sentiment model. Run the modified code using SAS® Studio.

Click the Hot Fix tab in this note for a link to instructions about accessing and applying the software update.

 

Example:

1. Create a test for a project that contains a customer-sentiment model in SAS Model Manager.

2. Run the test.

3. Click Results ► Code. Code like the following is displayed. 

          proc cas;
          session _mmcas_;
          loadactionset "sentimentAnalysis";
          action applySent;
          param
          table={caslib="&caslib_name", name="WINE_QUALITY_REVIEWS"}
          docId=&key_column
          text=&document_column
          language=&language
          casOut={caslib="&output_caslib_name", name="&output_table_name", replace=TRUE}
          matchOut={caslib="&output_caslib_name", name=&output_matches_table_name, replace=TRUE}
          featureOut={caslib="&output_caslib_name", name=&output_features_table_name, replace=TRUE}


               This code is not correct because it does not specify the customer-sentiment model. 
 

 

4. Replace the code above with this code:

          proc cas;
             session sascas1;
             loadactionset "sentimentAnalysis";
             action applySent;
                param
                   model={caslib=&sentiment_binary_caslib, name=&sentiment_binary_table_name}
                   table={caslib=&input_caslib_name, name=&input_table_name}
                   docId=&key_column
                   text=&document_column
                   casOut={caslib=&output_caslib_name, name=&output_sentiment_table_name, replace=TRUE}
                   matchOut={caslib=&output_caslib_name, name=&output_matches_table_name, replace=TRUE}
                   featureOut={caslib=&output_caslib_name, name=&output_features_table_name, replace=TRUE}

 

5. Run the modified code using SAS Studio.



Operating System and Release Information

Product FamilyProductSystemProduct ReleaseSAS Release
ReportedFixed*ReportedFixed*
SAS SystemSAS Model ManagerLinux for x6415.2Viya
* For software releases that are not yet generally available, the Fixed Release is the software release in which the problem is planned to be fixed.