[ML] calculate feature importance for multi-class results

[benwtrent]

Feature importance is already calculated for multi-class models. This commit adjusts the output sent to ES so that multi-class importance can be explored.

Feature importance objects are now mapped as follows
(logistic) Regression:

{
   "feature_name": "feature_0",
   "importance": -1.3
}

Multi-class [class names are foo, bar, baz]

{ 
   “feature_name”: “feature_0”, 
   “importance”: 2.0, // sum(abs()) of class importances
   “foo”: 1.0, 
   “bar”: 0.5, 
   “baz”: -0.5 
},

Java side change: elastic/elasticsearch#53803

[benwtrent]

Ugh, windows build failed due to

CDataFrameAnalyzerFeatureImportanceTest.obj : error LNK2019: unresolved external symbol "public: static class ml::maths::CDenseVector<double> __cdecl ml::maths::CTools::softmax<double>(class ml::maths::CDenseVector<double>)" (??$softmax@N@CTools@maths@ml@@SA?AV?$CDenseVector@N@12@V312@@Z) referenced in function "public: class ml::maths::CDenseVector<double> __cdecl <lambda_c8f0b938f55e935ee669e05079912d7a>::operator()(class std::vector<double,class std::allocator<double> > const &)const " (??R<lambda_c8f0b938f55e935ee669e05079912d7a>@@QEBA?AV?$CDenseVector@N@maths@ml@@AEBV?$vector@NV?$allocator@N@std@@@std@@@Z)
ml_test.exe : fatal error LNK1120: 1 unresolved externals

I suppose I need to update the make file or something?

Tests ran and passed locally without issues though :(

[tveasey]

Overall looks good Ben! Just some minor comments.

For the compilation failure you just need to add #include <maths/CToolsDetail.h> in CDataFrameAnalyzerFeatureImportanceTest.cc.

[tveasey]

LGTM