/**
* Validate {@link ClusterContingencyTable} with respect to its ability to compare data
* clusterings.
*
* @throws ParameterException on errors.
*/
@Test
public void testCompareDatabases() {
ListParameterization params = new ListParameterization();
// Input
params.addParameter(FileBasedDatabaseConnection.Parameterizer.INPUT_ID, dataset);
// get database
Database db = ClassGenericsUtil.parameterizeOrAbort(StaticArrayDatabase.class, params);
db.initialize();
// verify data set size.
Relation<?> rel = db.getRelation(TypeUtil.ANY);
assertTrue(rel.size() == shoulds);
// run all-in-one
TrivialAllInOne allinone = new TrivialAllInOne();
Clustering<Model> rai = allinone.run(db);
// run all-in-noise
TrivialAllNoise allinnoise = new TrivialAllNoise();
Clustering<Model> ran = allinnoise.run(db);
// run by-label
ByLabelClustering bylabel = new ByLabelClustering();
Clustering<?> rbl = bylabel.run(db);
assertEquals(1.0, computeFMeasure(rai, rai, false), Double.MIN_VALUE);
assertEquals(1.0, computeFMeasure(ran, ran, false), Double.MIN_VALUE);
assertEquals(1.0, computeFMeasure(rbl, rbl, false), Double.MIN_VALUE);
assertEquals(0.009950248756218905, computeFMeasure(ran, rbl, true), Double.MIN_VALUE);
assertEquals(0.0033277870216306157, computeFMeasure(rai, ran, true), Double.MIN_VALUE);
assertEquals(0.5 /* 0.3834296724470135 */, computeFMeasure(rai, rbl, false), Double.MIN_VALUE);
}
/**
* Run KMeans with fixed parameters and compare the result to a golden standard.
*
* @throws ParameterException
*/
@Test
public void testKMeansHybridLloydMacQueen() {
Database db = makeSimpleDatabase(UNITTEST + "different-densities-2d-no-noise.ascii", 1000);
// Setup algorithm
ListParameterization params = new ListParameterization();
params.addParameter(KMeans.K_ID, 5);
params.addParameter(KMeans.SEED_ID, 2);
AbstractKMeans<DoubleVector, ?> kmeans =
ClassGenericsUtil.parameterizeOrAbort(KMeansHybridLloydMacQueen.class, params);
testParameterizationOk(params);
// run KMeans on database
Clustering<?> result = kmeans.run(db);
testFMeasure(db, result, 0.998005);
testClusterSizes(result, new int[] {199, 200, 200, 200, 201});
}
@Test
public void testLOF() {
Database db = makeSimpleDatabase(UNITTEST + "outlier-axis-subspaces-6d.ascii", 1345);
// Parameterization
ListParameterization params = new ListParameterization();
params.addParameter(LOF.Parameterizer.K_ID, 10);
// setup Algorithm
LOF<DoubleVector> lof = ClassGenericsUtil.parameterizeOrAbort(LOF.class, params);
testParameterizationOk(params);
// run LOF on database
OutlierResult result = lof.run(db);
testSingleScore(result, 1293, 1.1945314199156365);
testAUC(db, "Noise", result, 0.8921680672268908);
}
@Test
public void testKNNWeightOutlier() {
Database db = makeSimpleDatabase(UNITTEST + "outlier-3d-3clusters.ascii", 960);
// Parameterization
ListParameterization params = new ListParameterization();
params.addParameter(KNNWeightOutlier.Parameterizer.K_ID, 4);
// setup Algorithm
KNNWeightOutlier<DoubleVector> knnWeightOutlier =
ClassGenericsUtil.parameterizeOrAbort(KNNWeightOutlier.class, params);
testParameterizationOk(params);
// run KNNWeightOutlier on database
OutlierResult result = knnWeightOutlier.run(db);
testSingleScore(result, 945, 2.384117261027324);
testAUC(db, "Noise", result, 0.9912777777777778);
}
@Test
public void testDBOutlierDetection() {
Database db = makeSimpleDatabase(UNITTEST + "outlier-fire.ascii", 1025);
// Parameterization
ListParameterization params = new ListParameterization();
params.addParameter(DBOutlierDetection.Parameterizer.D_ID, 0.175);
params.addParameter(DBOutlierDetection.Parameterizer.P_ID, 0.98);
// setup Algorithm
DBOutlierDetection<DoubleVector> dbOutlierDetection =
ClassGenericsUtil.parameterizeOrAbort(DBOutlierDetection.class, params);
testParameterizationOk(params);
// run DBOutlierDetection on database
OutlierResult result = dbOutlierDetection.run(db);
testSingleScore(result, 1025, 0.0);
testAUC(db, "Noise", result, 0.97487179);
}
@Override
protected void makeOptions(Parameterization config) {
super.makeOptions(config);
format = ClassGenericsUtil.parameterizeOrAbort(CSVReaderFormat.class, config);
}