def kmeans_grid_iris():
iris_h2o = h2o.import_file(path=pyunit_utils.locate("smalldata/iris/iris.csv"))
grid_space = pyunit_utils.make_random_grid_space(algo="km")
print("Grid space: {0}".format(grid_space))
print("Constructing grid of Kmeans models")
iris_grid = H2OGridSearch(H2OKMeansEstimator, hyper_params=grid_space)
iris_grid.train(x=list(range(4)), training_frame=iris_h2o)
print("Check cardinality of grid, that is, the correct number of models have been created...")
size_of_grid_space = 1
for v in list(grid_space.values()):
size_of_grid_space = size_of_grid_space * len(v)
actual_size = len(iris_grid)
assert size_of_grid_space == actual_size, "Expected size of grid to be {0}, but got {1}" \
"".format(size_of_grid_space,actual_size)
print("Duplicate-entries-in-grid-space check")
new_grid_space = copy.deepcopy(grid_space)
for name in list(grid_space.keys()):
new_grid_space[name] = grid_space[name] + grid_space[name]
print("The new search space: {0}".format(new_grid_space))
print("Constructing the new grid of glm models...")
iris_grid2 = H2OGridSearch(H2OKMeansEstimator, hyper_params=new_grid_space)
iris_grid2.train(x=list(range(4)), training_frame=iris_h2o)
actual_size2 = len(iris_grid2)
assert actual_size == actual_size2, "Expected duplicates to be ignored. Without dups grid size: {0}. With dups " \
"size: {1}".format(actual_size, actual_size2)
print("Check that the hyper_params that were passed to grid, were used to construct the models...")
for name in list(grid_space.keys()):
print(name)
pyunit_utils.expect_model_param(iris_grid, name, grid_space[name])
def grid_cars_GBM():
cars = h2o.import_file(path=pyunit_utils.locate("smalldata/junit/cars_20mpg.csv"))
r = cars[0].runif(seed=42)
train = cars[r > .2]
validation_scheme = random.randint(1,3) # 1:none, 2:cross-validation, 3:validation set
print "Validation scheme: {0}".format(validation_scheme)
if validation_scheme == 2:
nfolds = 2
print "Nfolds: 2"
if validation_scheme == 3: valid = cars[r <= .2]
grid_space = pyunit_utils.make_random_grid_space(algo="gbm")
print "Grid space: {0}".format(grid_space)
predictors = ["displacement","power","weight","acceleration","year"]
if grid_space['distribution'][0] == 'bernoulli': response_col = "economy_20mpg"
elif grid_space['distribution'][0] == 'gaussian': response_col = "economy"
else: response_col = "cylinders"
print "Predictors: {0}".format(predictors)
print "Response: {0}".format(response_col)
if grid_space['distribution'][0] in ['bernoulli', 'multinomial']:
print "Converting the response column to a factor..."
train[response_col] = train[response_col].asfactor()
if validation_scheme == 3: valid[response_col] = valid[response_col].asfactor()
print "Constructing the grid of gbm models..."
cars_gbm_grid = H2OGridSearch(H2OGradientBoostingEstimator, hyper_params=grid_space)
if validation_scheme == 1: cars_gbm_grid.train(x=predictors,y=response_col,training_frame=train)
elif validation_scheme == 2: cars_gbm_grid.train(x=predictors,y=response_col,training_frame=train,nfolds=nfolds)
else: cars_gbm_grid.train(x=predictors,y=response_col,training_frame=train,validation_frame=valid)
print "Performing various checks of the constructed grid..."
print "Check cardinality of grid, that is, the correct number of models have been created..."
size_of_grid_space = 1
for v in grid_space.values(): size_of_grid_space = size_of_grid_space * len(v)
actual_size = len(cars_gbm_grid)
assert size_of_grid_space == actual_size, "Expected size of grid to be {0}, but got {1}" \
"".format(size_of_grid_space,actual_size)
print "Duplicate-entries-in-grid-space check"
new_grid_space = copy.deepcopy(grid_space)
for name in grid_space.keys():
if not name == "distribution": new_grid_space[name] = grid_space[name] + grid_space[name]
print "The new search space: {0}".format(new_grid_space)
print "Constructing the new grid of gbm models..."
cars_gbm_grid2 = H2OGridSearch(H2OGradientBoostingEstimator, hyper_params=new_grid_space)
if validation_scheme == 1: cars_gbm_grid2.train(x=predictors,y=response_col,training_frame=train)
elif validation_scheme == 2: cars_gbm_grid2.train(x=predictors,y=response_col,training_frame=train,nfolds=nfolds)
else: cars_gbm_grid2.train(x=predictors,y=response_col,training_frame=train,validation_frame=valid)
actual_size2 = len(cars_gbm_grid2)
assert actual_size == actual_size2, "Expected duplicates to be ignored. Without dups grid size: {0}. With dups " \
"size: {1}".format(actual_size, actual_size2)
print "Check that the hyper_params that were passed to grid, were used to construct the models..."
for name in grid_space.keys(): pyunit_utils.expect_model_param(cars_gbm_grid, name, grid_space[name])
def kmeans_grid_iris():
iris_h2o = h2o.import_file(
path=pyunit_utils.locate("smalldata/iris/iris.csv"))
grid_space = pyunit_utils.make_random_grid_space(algo="km")
print "Grid space: {0}".format(grid_space)
print "Constructing grid of Kmeans models"
iris_grid = H2OGridSearch(H2OKMeansEstimator, hyper_params=grid_space)
iris_grid.train(x=range(4), training_frame=iris_h2o)
print "Check cardinality of grid, that is, the correct number of models have been created..."
size_of_grid_space = 1
for v in grid_space.values():
size_of_grid_space = size_of_grid_space * len(v)
actual_size = len(iris_grid)
assert size_of_grid_space == actual_size, "Expected size of grid to be {0}, but got {1}" \
"".format(size_of_grid_space,actual_size)
print "Duplicate-entries-in-grid-space check"
new_grid_space = copy.deepcopy(grid_space)
for name in grid_space.keys():
new_grid_space[name] = grid_space[name] + grid_space[name]
print "The new search space: {0}".format(new_grid_space)
print "Constructing the new grid of glm models..."
iris_grid2 = H2OGridSearch(H2OKMeansEstimator, hyper_params=new_grid_space)
iris_grid2.train(x=range(4), training_frame=iris_h2o)
actual_size2 = len(iris_grid2)
assert actual_size == actual_size2, "Expected duplicates to be ignored. Without dups grid size: {0}. With dups " \
"size: {1}".format(actual_size, actual_size2)
print "Check that the hyper_params that were passed to grid, were used to construct the models..."
for name in grid_space.keys():
print name
pyunit_utils.expect_model_param(iris_grid, name, grid_space[name])
def grid_quasar_pca():
quasar = h2o.import_file(
path=pyunit_utils.locate("smalldata/pca_test/SDSS_quasar.txt.zip"),
header=1)
grid_space = pyunit_utils.make_random_grid_space(algo="pca",
ncols=quasar.ncol,
nrows=quasar.nrow)
print("Grid space: {0}".format(grid_space))
print("Constructing the grid of PCA models...")
quasar_pca_grid = H2OGridSearch(H2OPCA, hyper_params=grid_space)
quasar_pca_grid.train(x=list(range(1, 23)), training_frame=quasar)
for model in quasar_pca_grid:
assert isinstance(model, H2OPCA)
print("Performing various checks of the constructed grid...")
print(
"Check cardinality of grid, that is, the correct number of models have been created..."
)
size_of_grid_space = 1
for v in list(grid_space.values()):
v2 = [v] if type(v) != list else v
size_of_grid_space = size_of_grid_space * len(v2)
actual_size = len(quasar_pca_grid)
assert size_of_grid_space == actual_size, "Expected size of grid to be {0}, but got {1}" \
"".format(size_of_grid_space,actual_size)
print("Duplicate-entries-in-grid-space check")
new_grid_space = copy.deepcopy(grid_space)
for name in list(grid_space.keys()):
new_grid_space[name] = grid_space[name] + grid_space[name]
print("The new search space: {0}".format(new_grid_space))
print("Constructing the new grid of nb models...")
quasar_pca_grid2 = H2OGridSearch(H2OPCA, hyper_params=new_grid_space)
quasar_pca_grid2.train(x=list(range(1, 23)), training_frame=quasar)
actual_size2 = len(quasar_pca_grid2)
assert actual_size == actual_size2, "Expected duplicates to be ignored. Without dups grid size: {0}. With dups " \
"size: {1}".format(actual_size, actual_size2)
for model in quasar_pca_grid2:
assert isinstance(model, H2OPCA)
print(
"Check that the hyper_params that were passed to grid, were used to construct the models..."
)
for name in list(grid_space.keys()):
print(name)
pyunit_utils.expect_model_param(quasar_pca_grid, name,
grid_space[name])
def grid_cars_NB():
cars = h2o.import_file(path=pyunit_utils.locate("smalldata/junit/cars_20mpg.csv"))
r = cars[0].runif(seed=42)
train = cars[r > .2]
validation_scheme = random.randint(1,3) # 1:none, 2:cross-validation, 3:validation set
print("Validation scheme: {0}".format(validation_scheme))
if validation_scheme == 2:
nfolds = 2
print("Nfolds: 2")
if validation_scheme == 3:
valid = cars[r <= .2]
grid_space = pyunit_utils.make_random_grid_space(algo="naiveBayes")
print("Grid space: {0}".format(grid_space))
problem = random.sample(["binomial","multinomial"],1)
predictors = ["displacement","power","weight","acceleration","year"]
if problem == "binomial":
response_col = "economy_20mpg"
else:
response_col = "cylinders"
print("Predictors: {0}".format(predictors))
print("Response: {0}".format(response_col))
print("Converting the response column to a factor...")
train[response_col] = train[response_col].asfactor()
if validation_scheme == 3:
valid[response_col] = valid[response_col].asfactor()
print("Grid space: {0}".format(grid_space))
print("Constructing the grid of nb models...")
grid_space["compute_metrics"] = [False]
cars_nb_grid = H2OGridSearch(H2ONaiveBayesEstimator, hyper_params=grid_space)
if validation_scheme == 1:
cars_nb_grid.train(x=predictors,y=response_col,training_frame=train)
elif validation_scheme == 2:
cars_nb_grid.train(x=predictors,y=response_col,training_frame=train,nfolds=nfolds)
else:
cars_nb_grid.train(x=predictors,y=response_col,training_frame=train,validation_frame=valid)
def grid_quasar_pca():
quasar = h2o.import_file(path=pyunit_utils.locate("smalldata/pca_test/SDSS_quasar.txt.zip"), header=1)
grid_space = pyunit_utils.make_random_grid_space(algo="pca", ncols=quasar.ncol, nrows=quasar.nrow)
print("Grid space: {0}".format(grid_space))
print("Constructing the grid of PCA models...")
quasar_pca_grid = H2OGridSearch(H2OPCA, hyper_params=grid_space)
quasar_pca_grid.train(x=list(range(1,23)), training_frame=quasar)
for model in quasar_pca_grid:
assert isinstance(model, H2OPCA)
print("Performing various checks of the constructed grid...")
print("Check cardinality of grid, that is, the correct number of models have been created...")
size_of_grid_space = 1
for v in list(grid_space.values()):
v2 = [v] if type(v) != list else v
size_of_grid_space = size_of_grid_space * len(v2)
actual_size = len(quasar_pca_grid)
assert size_of_grid_space == actual_size, "Expected size of grid to be {0}, but got {1}" \
"".format(size_of_grid_space,actual_size)
print("Duplicate-entries-in-grid-space check")
new_grid_space = copy.deepcopy(grid_space)
for name in list(grid_space.keys()):
new_grid_space[name] = grid_space[name] + grid_space[name]
print("The new search space: {0}".format(new_grid_space))
print("Constructing the new grid of nb models...")
quasar_pca_grid2 = H2OGridSearch(H2OPCA, hyper_params=new_grid_space)
quasar_pca_grid2.train(x=list(range(1,23)), training_frame=quasar)
actual_size2 = len(quasar_pca_grid2)
assert actual_size == actual_size2, "Expected duplicates to be ignored. Without dups grid size: {0}. With dups " \
"size: {1}".format(actual_size, actual_size2)
for model in quasar_pca_grid2:
assert isinstance(model, H2OPCA)
print("Check that the hyper_params that were passed to grid, were used to construct the models...")
for name in list(grid_space.keys()):
print(name)
pyunit_utils.expect_model_param(quasar_pca_grid, name, grid_space[name])
def grid_cars_RF():
cars = h2o.import_file(
path=pyunit_utils.locate("smalldata/junit/cars_20mpg.csv"))
r = cars[0].runif(seed=42)
train = cars[r > .2]
validation_scheme = random.randint(
1, 3) # 1:none, 2:cross-validation, 3:validation set
print("Validation scheme: {0}".format(validation_scheme))
if validation_scheme == 2:
nfolds = 2
print("Nfolds: 2")
if validation_scheme == 3:
valid = cars[r <= .2]
predictors = ["displacement", "power", "weight", "acceleration", "year"]
grid_space = pyunit_utils.make_random_grid_space(algo="rf",
ncols=len(predictors))
print("Grid space: {0}".format(grid_space))
problem = random.randint(1, 3)
if problem == 1:
response_col = "economy_20mpg"
elif problem == 2:
response_col = "economy"
else:
response_col = "cylinders"
print("Predictors: {0}".format(predictors))
print("Response: {0}".format(response_col))
if problem in [1, 3]:
print("Converting the response column to a factor...")
train[response_col] = train[response_col].asfactor()
if validation_scheme == 3:
valid[response_col] = valid[response_col].asfactor()
print("Constructing the grid of RF models...")
cars_rf_grid = H2OGridSearch(H2ORandomForestEstimator,
hyper_params=grid_space)
if validation_scheme == 1:
cars_rf_grid.train(x=predictors, y=response_col, training_frame=train)
elif validation_scheme == 2:
cars_rf_grid.train(x=predictors,
y=response_col,
training_frame=train,
nfolds=nfolds)
else:
cars_rf_grid.train(x=predictors,
y=response_col,
training_frame=train,
validation_frame=valid)
for model in cars_rf_grid:
assert isinstance(model, H2ORandomForestEstimator)
print("Performing various checks of the constructed grid...")
print(
"Check cardinality of grid, that is, the correct number of models have been created..."
)
size_of_grid_space = 1
for v in list(grid_space.values()):
size_of_grid_space = size_of_grid_space * len(v)
actual_size = len(cars_rf_grid)
assert size_of_grid_space == actual_size, "Expected size of grid to be {0}, but got {1}" \
"".format(size_of_grid_space,actual_size)
print("Duplicate-entries-in-grid-space check")
new_grid_space = copy.deepcopy(grid_space)
for name in list(grid_space.keys()):
if not name == "distribution":
new_grid_space[name] = grid_space[name] + grid_space[name]
print("The new search space: {0}".format(new_grid_space))
print("Constructing the new grid of RF models...")
cars_rf_grid2 = H2OGridSearch(H2ORandomForestEstimator,
hyper_params=new_grid_space)
if validation_scheme == 1:
cars_rf_grid2.train(x=predictors, y=response_col, training_frame=train)
elif validation_scheme == 2:
cars_rf_grid2.train(x=predictors,
y=response_col,
training_frame=train,
nfolds=nfolds)
else:
cars_rf_grid2.train(x=predictors,
y=response_col,
training_frame=train,
validation_frame=valid)
actual_size2 = len(cars_rf_grid2)
assert actual_size == actual_size2, "Expected duplicates to be ignored. Without dups grid size: {0}. With dups " \
"size: {1}".format(actual_size, actual_size2)
for model in cars_rf_grid2:
assert isinstance(model, H2ORandomForestEstimator)
print(grid_space)
print(
"Check that the hyper_params that were passed to grid, were used to construct the models..."
)
for name in list(grid_space.keys()):
pyunit_utils.expect_model_param(cars_rf_grid, name, grid_space[name])
def grid_cars_GLM():
cars = h2o.import_file(
path=pyunit_utils.locate("smalldata/junit/cars_20mpg.csv"))
r = cars[0].runif(seed=42)
train = cars[r > .2]
validation_scheme = random.randint(
1, 3) # 1:none, 2:cross-validation, 3:validation set
print("Validation scheme: {0}".format(validation_scheme))
if validation_scheme == 2:
nfolds = 2
print("Nfolds: 2")
if validation_scheme == 3:
valid = cars[r <= .2]
grid_space = pyunit_utils.make_random_grid_space(algo="glm")
print("Grid space: {0}".format(grid_space))
predictors = ["displacement", "power", "weight", "acceleration", "year"]
if grid_space['family'][0] == 'binomial':
response_col = "economy_20mpg"
true_model_type = "classifier"
elif grid_space['family'][0] == 'gaussian':
response_col = "economy"
true_model_type = "regressor"
else:
response_col = "cylinders"
true_model_type = "regressor"
print("Predictors: {0}".format(predictors))
print("Response: {0}".format(response_col))
if grid_space['family'][0] in ['binomial', 'multinomial']:
print("Converting the response column to a factor...")
train[response_col] = train[response_col].asfactor()
if validation_scheme == 3:
valid[response_col] = valid[response_col].asfactor()
#grid_space.update({"lambda":[0.1,0.05,0.01]})
family = grid_space.pop('family')[0]
print("Grid space: {0}".format(grid_space))
print("Constructing the grid of glm models...")
print("family = ", family)
cars_glm_grid = H2OGridSearch(H2OGeneralizedLinearEstimator,
hyper_params=grid_space)
if validation_scheme == 1:
cars_glm_grid.train(x=predictors,
y=response_col,
training_frame=train,
family=family)
elif validation_scheme == 2:
cars_glm_grid.train(x=predictors,
y=response_col,
training_frame=train,
nfolds=nfolds,
family=family)
else:
cars_glm_grid.train(x=predictors,
y=response_col,
training_frame=train,
validation_frame=valid,
family=family)
for model in cars_glm_grid:
assert isinstance(model, H2OGeneralizedLinearEstimator)
print("Performing various checks of the constructed grid...")
print(
"Check cardinality of grid, that is, the correct number of models have been created..."
)
size_of_grid_space = 1
for v in list(grid_space.values()):
size_of_grid_space = size_of_grid_space * len(v)
actual_size = len(cars_glm_grid)
assert size_of_grid_space == actual_size, "Expected size of grid to be {0}, but got {1}" \
"".format(size_of_grid_space,actual_size)
print("Check correct type value....")
model_type = cars_glm_grid[0].type
assert model_type == true_model_type, "Type of model ({0}) is incorrect, expected value is {1}.".format(
model_type, true_model_type)
print("Duplicate-entries-in-grid-space check")
new_grid_space = copy.deepcopy(grid_space)
for name in list(grid_space.keys()):
if not name == "family":
new_grid_space[name] = grid_space[name] + grid_space[name]
print("The new search space: {0}".format(new_grid_space))
print("Constructing the new grid of glm models...")
cars_glm_grid2 = H2OGridSearch(H2OGeneralizedLinearEstimator,
hyper_params=new_grid_space)
if validation_scheme == 1:
cars_glm_grid2.train(x=predictors,
y=response_col,
training_frame=train,
family=family)
elif validation_scheme == 2:
cars_glm_grid2.train(x=predictors,
y=response_col,
training_frame=train,
nfolds=nfolds,
family=family)
else:
cars_glm_grid2.train(x=predictors,
y=response_col,
training_frame=train,
validation_frame=valid,
family=family)
actual_size2 = len(cars_glm_grid2)
assert actual_size == actual_size2, "Expected duplicates to be ignored. Without dups grid size: {0}. With dups " \
"size: {1}".format(actual_size, actual_size2)
print(
"Check that the hyper_params that were passed to grid, were used to construct the models..."
)
for name in list(grid_space.keys()):
print(name)
pyunit_utils.expect_model_param(cars_glm_grid, name, grid_space[name])
def grid_cars_GLM():
cars = h2o.import_file(path=pyunit_utils.locate("smalldata/junit/cars_20mpg.csv"))
r = cars[0].runif(seed=42)
train = cars[r > .2]
validation_scheme = random.randint(1,3) # 1:none, 2:cross-validation, 3:validation set
print("Validation scheme: {0}".format(validation_scheme))
if validation_scheme == 2:
nfolds = 2
print("Nfolds: 2")
if validation_scheme == 3:
valid = cars[r <= .2]
grid_space = pyunit_utils.make_random_grid_space(algo="glm")
print("Grid space: {0}".format(grid_space))
predictors = ["displacement","power","weight","acceleration","year"]
if grid_space['family'][0] == 'binomial':
response_col = "economy_20mpg"
true_model_type = "classifier"
elif grid_space['family'][0] == 'gaussian':
response_col = "economy"
true_model_type = "regressor"
else:
response_col = "cylinders"
true_model_type = "regressor"
print("Predictors: {0}".format(predictors))
print("Response: {0}".format(response_col))
if grid_space['family'][0] in ['binomial', 'multinomial']:
print("Converting the response column to a factor...")
train[response_col] = train[response_col].asfactor()
if validation_scheme == 3:
valid[response_col] = valid[response_col].asfactor()
#grid_space.update({"lambda":[0.1,0.05,0.01]})
family = grid_space.pop('family')[0]
print("Grid space: {0}".format(grid_space))
print("Constructing the grid of glm models...")
print("family = ",family)
cars_glm_grid = H2OGridSearch(H2OGeneralizedLinearEstimator, hyper_params=grid_space)
if validation_scheme == 1:
cars_glm_grid.train(x=predictors,y=response_col,training_frame=train, family=family)
elif validation_scheme == 2:
cars_glm_grid.train(x=predictors,y=response_col,training_frame=train,nfolds=nfolds, family=family)
else:
cars_glm_grid.train(x=predictors,y=response_col,training_frame=train,validation_frame=valid, family=family)
for model in cars_glm_grid:
assert isinstance(model, H2OGeneralizedLinearEstimator)
print("Performing various checks of the constructed grid...")
print("Check cardinality of grid, that is, the correct number of models have been created...")
size_of_grid_space = 1
for v in list(grid_space.values()):
size_of_grid_space = size_of_grid_space * len(v)
actual_size = len(cars_glm_grid)
assert size_of_grid_space == actual_size, "Expected size of grid to be {0}, but got {1}" \
"".format(size_of_grid_space,actual_size)
print("Check correct type value....")
model_type = cars_glm_grid[0].type
assert model_type == true_model_type, "Type of model ({0}) is incorrect, expected value is {1}.".format(model_type, true_model_type)
print("Duplicate-entries-in-grid-space check")
new_grid_space = copy.deepcopy(grid_space)
for name in list(grid_space.keys()):
if not name == "family":
new_grid_space[name] = grid_space[name] + grid_space[name]
print("The new search space: {0}".format(new_grid_space))
print("Constructing the new grid of glm models...")
cars_glm_grid2 = H2OGridSearch(H2OGeneralizedLinearEstimator, hyper_params=new_grid_space)
if validation_scheme == 1:
cars_glm_grid2.train(x=predictors,y=response_col,training_frame=train, family=family)
elif validation_scheme == 2:
cars_glm_grid2.train(x=predictors,y=response_col,training_frame=train,nfolds=nfolds, family=family)
else:
cars_glm_grid2.train(x=predictors,y=response_col,training_frame=train,validation_frame=valid, family=family)
actual_size2 = len(cars_glm_grid2)
assert actual_size == actual_size2, "Expected duplicates to be ignored. Without dups grid size: {0}. With dups " \
"size: {1}".format(actual_size, actual_size2)
print("Check that the hyper_params that were passed to grid, were used to construct the models...")
for name in list(grid_space.keys()):
print(name)
pyunit_utils.expect_model_param(cars_glm_grid, name, grid_space[name])
def grid_cars_RF():
cars = h2o.import_file(path=pyunit_utils.locate("smalldata/junit/cars_20mpg.csv"))
r = cars[0].runif(seed=42)
train = cars[r > .2]
validation_scheme = random.randint(1,3) # 1:none, 2:cross-validation, 3:validation set
print("Validation scheme: {0}".format(validation_scheme))
if validation_scheme == 2:
nfolds = 2
print("Nfolds: 2")
if validation_scheme == 3:
valid = cars[r <= .2]
predictors = ["displacement","power","weight","acceleration","year"]
grid_space = pyunit_utils.make_random_grid_space(algo="rf", ncols=len(predictors))
# reduce the magnitude of nbins_cats, run was too long.
if 'nbins_cats' in list(grid_space):
grid_space['nbins_cats'] = random.sample(list(range(2, 200)), random.randint(2, 3))
print("Grid space: {0}".format(grid_space))
problem = random.randint(1,3)
if problem == 1:
response_col = "economy_20mpg"
true_model_type = "classifier"
elif problem == 2:
response_col = "economy"
true_model_type = "regressor"
else:
response_col = "cylinders"
true_model_type = "classifier"
print("Predictors: {0}".format(predictors))
print("Response: {0}".format(response_col))
if problem in [1,3]:
print("Converting the response column to a factor...")
train[response_col] = train[response_col].asfactor()
if validation_scheme == 3:
valid[response_col] = valid[response_col].asfactor()
print("Constructing the grid of RF models...")
cars_rf_grid = H2OGridSearch(H2ORandomForestEstimator, hyper_params=grid_space)
if validation_scheme == 1:
cars_rf_grid.train(x=predictors,y=response_col, training_frame=train)
elif validation_scheme == 2:
cars_rf_grid.train(x=predictors,y=response_col, training_frame=train, nfolds=nfolds)
else:
cars_rf_grid.train(x=predictors,y=response_col, training_frame=train, validation_frame=valid)
for model in cars_rf_grid:
assert isinstance(model, H2ORandomForestEstimator)
print("Performing various checks of the constructed grid...")
print("Check cardinality of grid, that is, the correct number of models have been created...")
size_of_grid_space = 1
for v in list(grid_space.values()):
size_of_grid_space = size_of_grid_space * len(v)
actual_size = len(cars_rf_grid)
assert size_of_grid_space == actual_size, "Expected size of grid to be {0}, but got {1}" \
"".format(size_of_grid_space,actual_size)
print("Check correct type value....")
model_type = cars_rf_grid[0].type
assert model_type == true_model_type, "Type of model ({0}) is incorrect, expected value is {1}.".format(model_type, true_model_type)
print("Duplicate-entries-in-grid-space check")
new_grid_space = copy.deepcopy(grid_space)
for name in list(grid_space.keys()):
if not name == "distribution":
new_grid_space[name] = grid_space[name] + grid_space[name]
print("The new search space: {0}".format(new_grid_space))
print("Constructing the new grid of RF models...")
cars_rf_grid2 = H2OGridSearch(H2ORandomForestEstimator, hyper_params=new_grid_space)
if validation_scheme == 1:
cars_rf_grid2.train(x=predictors,y=response_col, training_frame=train)
elif validation_scheme == 2:
cars_rf_grid2.train(x=predictors,y=response_col, training_frame=train, nfolds=nfolds)
else:
cars_rf_grid2.train(x=predictors,y=response_col, training_frame=train, validation_frame=valid)
actual_size2 = len(cars_rf_grid2)
assert actual_size == actual_size2, "Expected duplicates to be ignored. Without dups grid size: {0}. With dups " \
"size: {1}".format(actual_size, actual_size2)
for model in cars_rf_grid2:
assert isinstance(model, H2ORandomForestEstimator)
print(grid_space)
print("Check that the hyper_params that were passed to grid, were used to construct the models...")
for name in list(grid_space.keys()):
pyunit_utils.expect_model_param(cars_rf_grid, name, grid_space[name])
def grid_cars_NB():
cars = h2o.import_file(path=pyunit_utils.locate("smalldata/junit/cars_20mpg.csv"))
r = cars[0].runif(seed=42)
train = cars[r > .2]
validation_scheme = random.randint(1,3) # 1:none, 2:cross-validation, 3:validation set
print("Validation scheme: {0}".format(validation_scheme))
if validation_scheme == 2:
nfolds = 2
print("Nfolds: 2")
if validation_scheme == 3:
valid = cars[r <= .2]
grid_space = pyunit_utils.make_random_grid_space(algo="naiveBayes")
print("Grid space: {0}".format(grid_space))
problem = random.sample(["binomial","multinomial"],1)
predictors = ["displacement","power","weight","acceleration","year"]
if problem == "binomial":
response_col = "economy_20mpg"
else:
response_col = "cylinders"
print("Predictors: {0}".format(predictors))
print("Response: {0}".format(response_col))
print("Converting the response column to a factor...")
train[response_col] = train[response_col].asfactor()
if validation_scheme == 3:
valid[response_col] = valid[response_col].asfactor()
print("Grid space: {0}".format(grid_space))
print("Constructing the grid of nb models...")
cars_nb_grid = H2OGridSearch(H2ONaiveBayesEstimator, hyper_params=grid_space)
if validation_scheme == 1:
cars_nb_grid.train(x=predictors,y=response_col,training_frame=train)
elif validation_scheme == 2:
cars_nb_grid.train(x=predictors,y=response_col,training_frame=train,nfolds=nfolds)
else:
cars_nb_grid.train(x=predictors,y=response_col,training_frame=train,validation_frame=valid)
for model in cars_nb_grid:
assert isinstance(model, H2ONaiveBayesEstimator)
print("Performing various checks of the constructed grid...")
print("Check cardinality of grid, that is, the correct number of models have been created...")
size_of_grid_space = 1
print(grid_space)
for v in list(grid_space.values()):
v2 = [v] if type(v) != list else v
size_of_grid_space = size_of_grid_space * len(v2)
actual_size = len(cars_nb_grid)
assert size_of_grid_space == actual_size, "Expected size of grid to be {0}, but got {1}" \
"".format(size_of_grid_space,actual_size)
print("Duplicate-entries-in-grid-space check")
new_grid_space = copy.deepcopy(grid_space)
for name in list(grid_space.keys()):
new_grid_space[name] = grid_space[name] + grid_space[name]
print("The new search space: {0}".format(new_grid_space))
print("Constructing the new grid of nb models...")
cars_nb_grid2 = H2OGridSearch(H2ONaiveBayesEstimator, hyper_params=new_grid_space)
if validation_scheme == 1:
cars_nb_grid2.train(x=predictors,y=response_col,training_frame=train)
elif validation_scheme == 2:
cars_nb_grid2.train(x=predictors,y=response_col,training_frame=train,nfolds=nfolds)
else:
cars_nb_grid2.train(x=predictors,y=response_col,training_frame=train,validation_frame=valid)
actual_size2 = len(cars_nb_grid2)
assert actual_size == actual_size2, "Expected duplicates to be ignored. Without dups grid size: {0}. With dups " \
"size: {1}".format(actual_size, actual_size2)
for model in cars_nb_grid2:
assert isinstance(model, H2ONaiveBayesEstimator)
print("Check that the hyper_params that were passed to grid, were used to construct the models...")
for name in list(grid_space.keys()):
print(name)
pyunit_utils.expect_model_param(cars_nb_grid, name, grid_space[name])
def grid_cars_NB():
cars = h2o.import_file(
path=pyunit_utils.locate("smalldata/junit/cars_20mpg.csv"))
r = cars[0].runif(seed=42)
train = cars[r > .2]
validation_scheme = random.randint(
1, 3) # 1:none, 2:cross-validation, 3:validation set
print "Validation scheme: {0}".format(validation_scheme)
if validation_scheme == 2:
nfolds = 2
print "Nfolds: 2"
if validation_scheme == 3:
valid = cars[r <= .2]
grid_space = pyunit_utils.make_random_grid_space(algo="naiveBayes")
print "Grid space: {0}".format(grid_space)
problem = random.sample(["binomial", "multinomial"], 1)
predictors = ["displacement", "power", "weight", "acceleration", "year"]
if problem == "binomial":
response_col = "economy_20mpg"
else:
response_col = "cylinders"
print "Predictors: {0}".format(predictors)
print "Response: {0}".format(response_col)
print "Converting the response column to a factor..."
train[response_col] = train[response_col].asfactor()
if validation_scheme == 3:
valid[response_col] = valid[response_col].asfactor()
print "Grid space: {0}".format(grid_space)
print "Constructing the grid of nb models..."
cars_nb_grid = H2OGridSearch(H2ONaiveBayesEstimator,
hyper_params=grid_space)
if validation_scheme == 1:
cars_nb_grid.train(x=predictors, y=response_col, training_frame=train)
elif validation_scheme == 2:
cars_nb_grid.train(x=predictors,
y=response_col,
training_frame=train,
nfolds=nfolds)
else:
cars_nb_grid.train(x=predictors,
y=response_col,
training_frame=train,
validation_frame=valid)
print "Performing various checks of the constructed grid..."
print "Check cardinality of grid, that is, the correct number of models have been created..."
size_of_grid_space = 1
print grid_space
for v in grid_space.values():
v2 = [v] if type(v) != list else v
size_of_grid_space = size_of_grid_space * len(v2)
actual_size = len(cars_nb_grid)
assert size_of_grid_space == actual_size, "Expected size of grid to be {0}, but got {1}" \
"".format(size_of_grid_space,actual_size)
print "Duplicate-entries-in-grid-space check"
new_grid_space = copy.deepcopy(grid_space)
for name in grid_space.keys():
new_grid_space[name] = grid_space[name] + grid_space[name]
print "The new search space: {0}".format(new_grid_space)
print "Constructing the new grid of nb models..."
cars_nb_grid2 = H2OGridSearch(H2ONaiveBayesEstimator,
hyper_params=new_grid_space)
if validation_scheme == 1:
cars_nb_grid2.train(x=predictors, y=response_col, training_frame=train)
elif validation_scheme == 2:
cars_nb_grid2.train(x=predictors,
y=response_col,
training_frame=train,
nfolds=nfolds)
else:
cars_nb_grid2.train(x=predictors,
y=response_col,
training_frame=train,
validation_frame=valid)
actual_size2 = len(cars_nb_grid2)
assert actual_size == actual_size2, "Expected duplicates to be ignored. Without dups grid size: {0}. With dups " \
"size: {1}".format(actual_size, actual_size2)
print "Check that the hyper_params that were passed to grid, were used to construct the models..."
for name in grid_space.keys():
print name
pyunit_utils.expect_model_param(cars_nb_grid, name, grid_space[name])