Building a Recommendation Engine: The XELOPES Library - Realtime Data Mining - page 255

MiningAlgorithm algorithm ¼ GeneralUtils. createMiningAl-
gorithmInstance (className);
// Put it all together:
algorithm.setMiningInputStream( inputData );
algorithm.setOuterMiningTransform( mta );
algorithm.setMiningSettings( miningSettings );
algorithm.setMiningAlgorithmSpecification(

miningAlgor-

ithmSpecification );
algorithm.verify();

// Build the mining model:
MiningModel model ¼ algorithm.buildModel();
System. out .println("calc. time[s]: " + algorithm.

getTimeSpentToBuildModel());

// Write to PMML:
FileWriter

¼

writer new FileWriter("data/pmml/

SparseGridsModel.xml");
model.writePmml(writer);
We now apply the sparse grid model of the PMML classifier SparseGridsModel.
xml , created before, to a CSV file TEST_FILE and calculate the classification
rate. Before applying the classifier, the normalization taken from the model is
carried out:

// Read SG model from PMML file:
SparseGridsMiningModel model ¼ new SparseGridsMiningModel
();
FileReader

¼

reader new FileReader("data/pmml/

SparseGridsModel.xml");
model.readPmml(reader);
MiningAttribute modelTargetAttribute

¼ ((Supervised

MiningSettings)
model.getMiningSettings()).getTarget();
System.out.println("---- > PMML model read successfully");

// Open data source and transform into model format:
MiningInputStream

inputData0 ¼ new MiningCsvStream(

TEST_FILE );
MiningInputStream inputData

¼ model.transformIntoMo-

delFormat(inputData0);

// Get input metadata:
MiningDataSpecification inputMetaData

¼ inputData.

getMetaData();
CategoricalAttribute

¼

inputTargetAttribute

(CategoricalAttribute)