def testInitialBiasAndWeightsAffectResult(prostateDataset):
[traningDataset, testingDataset] = prostateDataset.randomSplit([0.9, 0.1],
1)
def createInitialDeepLearningDefinition():
return H2ODeepLearning(seed=42,
reproducible=True,
labelCol="CAPSULE",
featuresCols=["AGE", "RACE", "DPROS", "DCAPS"],
hidden=[
3,
])
referenceDeepLearning = createInitialDeepLearningDefinition()
referenceModel = referenceDeepLearning.fit(traningDataset)
referenceResult = referenceModel.transform(testingDataset)
deepLearning = createInitialDeepLearningDefinition()
matrix0 = DenseMatrix(3, 4,
[.1, .2, .3, .4, .4, .5, .6, .7, .7, .8, .9, .6],
False)
matrix1 = DenseMatrix(1, 3, [.2, .3, .4], False)
deepLearning.setInitialWeights([matrix0, matrix1])
deepLearning.setInitialBiases(
[DenseVector([.1, .2, .3]),
DenseVector([.1])])
model = deepLearning.fit(traningDataset)
result = model.transform(testingDataset)
unit_test_utils.assert_data_frames_have_different_values(
referenceResult, result)
def testBetaConstraintsAffectResult(spark, prostateDataset):
[traningDataset, testingDataset] = prostateDataset.randomSplit([0.9, 0.1],
1)
featuresCols = ["DPROS", "DCAPS", "RACE", "GLEASON"]
def createInitialGamDefinition():
return H2OGAM(featuresCols=featuresCols,
labelCol="CAPSULE",
seed=1,
splitRatio=0.8,
gamCols=["PSA", "AGE"])
referenceGam = createInitialGamDefinition()
referenceModel = referenceGam.fit(traningDataset)
referenceResult = referenceModel.transform(testingDataset)
betaConstraints = map(lambda feature: (feature, -1000, 1000, 1, 0.2),
featuresCols)
betaConstraintsFrame = spark.createDataFrame(
betaConstraints,
['names', 'lower_bounds', 'upper_bounds', 'beta_given', 'rho'])
gam = createInitialGamDefinition()
gam.setBetaConstraints(betaConstraintsFrame)
model = gam.fit(traningDataset)
result = model.transform(testingDataset)
unit_test_utils.assert_data_frames_have_different_values(
referenceResult, result)
def testPlugValuesAffectResult(spark, carsDatasetPath):
carsDataset=spark.read.csv(carsDatasetPath, header=True, inferSchema=True)
carsDataset=carsDataset.withColumn("economy_20mpg", carsDataset.economy_20mpg.cast("string"))
[traningDataset, testingDataset] = carsDataset.randomSplit([0.9, 0.1], 1)
def createInitialGlmDefinition():
featuresCols=["economy","displacement", "power", "weight", "acceleration", "year", "economy_20mpg"]
return H2OGLM(featuresCols=featuresCols, labelCol="cylinders", seed=1,splitRatio=0.8)
referenceGlm = createInitialGlmDefinition()
referenceModel = referenceGlm.fit(traningDataset)
referenceResult = referenceModel.transform(testingDataset)
plugValues = {
"economy": 1.1,
"displacement": 2.2,
"power": 3.3,
"weight": 4.4,
"acceleration": 5.5,
"year": 2000,
"economy_20mpg": "0"}
glm = createInitialGlmDefinition()
glm.setMissingValuesHandling("PlugValues")
glm.setPlugValues(plugValues)
model = glm.fit(traningDataset)
result = model.transform(testingDataset)
unit_test_utils.assert_data_frames_have_different_values(referenceResult, result)
def testH2OAutoMLClassifierBehavesDiffenrentlyThanH2OAutoMLRegressor(prostateDataset):
[trainingDateset, testingDataset] = prostateDataset.randomSplit([0.9, 0.1], 42)
regressor = setParametersForTesting(H2OAutoMLRegressor())
regressionModel = regressor.fit(trainingDateset)
regressionDataset = regressionModel.transform(testingDataset).drop("detailed_prediction")
classifier = setParametersForTesting(H2OAutoMLClassifier())
classificationModel = classifier.fit(trainingDateset)
classificationDataset = classificationModel.transform(testingDataset).drop("detailed_prediction")
unit_test_utils.assert_data_frames_have_different_values(regressionDataset, classificationDataset)
def testGridSearchWithDRFClassifierBehavesDiffenrentlyThanGridSearchWithDRFRegressor(
prostateDataset):
[trainingDateset, testingDataset] = prostateDataset.randomSplit([0.9, 0.1],
42)
regressor = createGridForProblemSpecificTesting(H2ODRFRegressor())
regressionModel = regressor.fit(trainingDateset)
regressionDataset = regressionModel.transform(testingDataset).drop(
"detailed_prediction")
classifier = createGridForProblemSpecificTesting(H2ODRFClassifier())
classificationModel = classifier.fit(trainingDateset)
classificationDataset = classificationModel.transform(testingDataset).drop(
"detailed_prediction")
unit_test_utils.assert_data_frames_have_different_values(
regressionDataset, classificationDataset)
def testCalibrationDataFrameCauseGenerationOfCalibratedProbabilities(prostateDataset):
prostateDataset = prostateDataset.withColumn("CAPSULE", prostateDataset.CAPSULE.cast("string"))
[trainingDataset, testingDataset, calibrationDataset] = prostateDataset.randomSplit([0.9, 0.05, 0.05], 1)
algo = H2ODRF(featuresCols=["AGE", "RACE", "DPROS", "DCAPS", "PSA", "VOL", "GLEASON"],
labelCol="CAPSULE",
seed=1,
splitRatio=0.8,
withDetailedPredictionCol=True,
calibrateModel=True,
calibrationDataFrame=calibrationDataset)
model = algo.fit(trainingDataset)
result = model.transform(testingDataset).cache()
probabilities = result.select("ID", "detailed_prediction.probabilities.0", "detailed_prediction.probabilities.1")
calibrated = result.select(
"ID",
"detailed_prediction.calibratedProbabilities.0",
"detailed_prediction.calibratedProbabilities.1")
unit_test_utils.assert_data_frames_have_different_values(probabilities, calibrated)
def testInteractionConstraintsAffectResult(spark, prostateDataset):
[traningDataset, testingDataset] = prostateDataset.randomSplit([0.9, 0.1],
1)
featureCols = ["AGE", "RACE", "DPROS", "DCAPS", "PSA", "VOL", "GLEASON"]
def createInitialXGBoostDefinition():
return H2OXGBoost(featuresCols=featureCols,
labelCol="CAPSULE",
seed=1,
splitRatio=0.8)
referenceXGBoost = createInitialXGBoostDefinition()
referenceModel = referenceXGBoost.fit(traningDataset)
referenceResult = referenceModel.transform(testingDataset)
xgboost = createInitialXGBoostDefinition()
xgboost.setInteractionConstraints([["DPROS", "DCAPS"],
["PSA", "VOL", "GLEASON"]])
model = xgboost.fit(traningDataset)
result = model.transform(testingDataset)
unit_test_utils.assert_data_frames_have_different_values(
referenceResult, result)
def _fit(self, dataset):
unit_test_utils.assert_data_frames_are_identical(expected, dataset)
unit_test_utils.assert_data_frames_have_different_values(unexpected, dataset)
return DummyTransformer()
