Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

Code Block
languagejava
// User code

TableSourceOp edges, nodeLabels; // used for training, can be constructed from Table
TableSourceOp nodeIds; // used for prediction, can be constructed from Table
AlgoOperator modelOp = new MetaPath2vecTrainOp()
      .linkFrom(edges, nodeLabels);
AlgoOperator result = new MetaPath2vecPredictOp()
      .linkFrom(modelOp, nodeIds);

// Developer code

public class MetaPath2vecTrainOp extends AlgoOperator<MetaPath2vecTrainOp> {

       @Internal 
       @Override
       public abstract Table[] compute(Table… inputs) {
               // do the computation logic here and return the graph embedding models
       }

}

public class MetaPath2VecPredictOp extends AlgoOperator<MetaPath2VecPredictOp> {
       
       @Internal 
       @Override
        public abstract Table[] compute(Table… inputs) {
               // do the computation logic here and return the graph embedding models
        }

}

Express the algorithms that are used/covered by machine learning libraries but usually not implemented as a transformer/estimator.

...

Code Block
languagejava
// User code:

TableSourceOp input;
AlgoOperator auc = new EvaluationAUCOp().linkFrom(input);

// Developer Code:

public class EvaluationAUCOp extends AlgoOperator<EvaluationAUCOp> {
       @Internal 
       @Override
        public abstract Table[] compute(Table… inputs) {
               // do the computation logic     
        }
}



Compatibility, Deprecation, and Migration Plan

...