class Hyperparams(hyperparams.Hyperparams):
# search over these hyperparameters to tune performance
q = hyperparams.UniformInt(
default=3,
lower=2,
upper=10,
description="degree of the polynomial to be fit",
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
r = hyperparams.UniformInt(
default=5,
lower=2,
upper=30,
description="rank of the coefficient tensors to be fit",
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
gamma = hyperparams.LogUniform(
default=.01,
lower=.0001,
upper=10,
description="l2 regularization to use on the tensor low-rank factors",
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
alpha = hyperparams.LogUniform(
default=.1,
lower=.001,
upper=1,
description="variance of the random initialization of the factors",
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
epochs = hyperparams.UniformInt(
default=30,
lower=1,
upper=100,
description="maximum iterations of LBFGS, or number of epochs of SFO",
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
# control parameters determined once during pipeline building then fixed
solver = hyperparams.Enumeration[str](
default="LBFGS",
values=["SFO", "LBFGS"],
description=
"solver to use: LBFGS better for small enough datasets, SFO does minibached stochastic quasi-Newton to scale to large dataset",
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/ControlParameter'
])
preprocess = hyperparams.Enumeration[str](
default="YES",
values=["YES", "NO"],
description=
"whether to use a preprocessing that tends to work well for tensor machines",
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/ControlParameter'
])
class Hyperparams(hyperparams.Hyperparams):
# control parameters determined once during pipeline building then fixed
coresetmultiplier = hyperparams.UniformInt(
default=4,
lower=2,
upper=7,
description=
"coreset size, as a multiple of the number of input features",
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
eps = hyperparams.LogUniform(
default=1e-6,
lower=1e-14,
upper=1e-2,
description="relative error stopping tolerance for IRLS solver",
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/ControlParameter'
])
maxIters = hyperparams.UniformInt(
default=100,
lower=50,
upper=500,
description="maximum iterations of IRLS",
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
class Hyperparams(hyperparams.Hyperparams):
shapelet_length = hyperparams.LogUniform(lower = 0, upper = 1, default = 0.1,
upper_inclusive = False, semantic_types = [
'https://metadata.datadrivendiscovery.org/types/ControlParameter'],
description = 'base shapelet length, expressed as fraction of length of time series')
num_shapelet_lengths = hyperparams.UniformInt(lower = 1, upper = 100, default = 2, semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'],
description = 'number of different shapelet lengths')
# default epoch size from https://tslearn.readthedocs.io/en/latest/auto_examples/plot_shapelets.html#sphx-glr-auto-examples-plot-shapelets-py
epochs = hyperparams.UniformInt(lower = 1, upper = sys.maxsize, default = 200, semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'],
description = 'number of training epochs')
pass
class Hyperparams(hyperparams.Hyperparams):
# search over these hyperparameters to tune performance
lparam = hyperparams.LogUniform(default=.01, lower=.0001, upper=1000,
description="l2 regularization to use on the regression",
semantic_types=['https://metadata.datadrivendiscovery.org/types/TuningParameter'])
degree = hyperparams.UniformInt(default=3, lower=2, upper=9,
description="degree of the polynomial to fit",
semantic_types=['https://metadata.datadrivendiscovery.org/types/TuningParameter'])
offset = hyperparams.LogUniform(default=.1, lower=.001, upper=2,
description="value of constant feature to use in the regression",
semantic_types=['https://metadata.datadrivendiscovery.org/types/TuningParameter'])
sf = hyperparams.LogUniform(default=.01, lower=.00001, upper=2,
description="scale factor to use in the regression",
semantic_types=['https://metadata.datadrivendiscovery.org/types/TuningParameter'])
# control parameters determined once during pipeline building then fixed
eps = hyperparams.LogUniform(default=1e-3, lower=1e-14, upper=1e-2,
description="relative error stopping tolerance for PCG solver",
semantic_types=['https://metadata.datadrivendiscovery.org/types/ControlParameter'])
maxIters = hyperparams.UniformInt(default=200, lower=50, upper=500,
description="maximum iterations of PCG",
semantic_types=['https://metadata.datadrivendiscovery.org/types/TuningParameter'])
class Hyperparams(hyperparams.Hyperparams):
# search over these hyperparameters to tune performance
lparam = hyperparams.LogUniform(
default=.01,
lower=.0001,
upper=1000,
description="l2 regularization to use for the kernel regression",
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
sigma = hyperparams.LogUniform(
default=.01,
lower=.0001,
upper=1000,
description="bandwidth (sigma) parameter for the kernel regression",
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
# control parameters determined once during pipeline building then fixed
eps = hyperparams.LogUniform(
default=1e-4,
lower=1e-14,
upper=1e-2,
description="relative error stopping tolerance for PCG solver",
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/ControlParameter'
])
maxIters = hyperparams.UniformInt(
default=200,
lower=50,
upper=500,
description="maximum iterations of PCG",
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
class SSC_OMPHyperparams(hyperparams.Hyperparams):
n_clusters = hyperparams.Bounded[int](lower=2,
upper=None,
default=2,
semantic_types=['https://metadata.datadrivendiscovery.org/types/ControlParameter'],
description="number of subspaces/clusters to learn")
sparsity_level = hyperparams.UniformInt(lower=3,
upper=50,
default=3,
upper_inclusive = True,
semantic_types=['https://metadata.datadrivendiscovery.org/types/TuningParameter'],
description="The algorithm terminates when it has selected this many regression coefficients.")
thresh = hyperparams.LogUniform(lower=1e-10,
upper=1,
default=1e-6,
upper_inclusive = True,
semantic_types=['https://metadata.datadrivendiscovery.org/types/TuningParameter'],
description="Minimum regression residual for termination.")
class RNNHyperparams(hyperparams.Hyperparams):
n_batch = hyperparams.Hyperparameter[int](
default=1,
description='Maximum number of iterations. Default is 300 ',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
n_max_epoch = hyperparams.Hyperparameter[int](
default=1000,
description='Maximum number of iterations. Default is 300 ',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
n_max_epoch_total = hyperparams.Hyperparameter[int](
default=100,
description='Maximum number of iterations. Default is 300 ',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
n_neurons = hyperparams.Hyperparameter[int](
default=256,
description='Maximum number of iterations. Default is 300 ',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
n_input_dim = hyperparams.Hyperparameter[int](
default=1,
description='Maximum number of iterations. Default is 300 ',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
n_dense_dim = hyperparams.Hyperparameter[int](
default=128,
description='Maximum number of iterations. Default is 300 ',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
n_output_dim = hyperparams.Hyperparameter[int](
default=3,
description='Maximum number of iterations. Default is 300 ',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
n_patience = hyperparams.Hyperparameter[int](
default=100,
description='Maximum number of iterations. Default is 300 ',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
n_lr_decay = hyperparams.Hyperparameter[int](
default=5,
description='Maximum number of iterations. Default is 300 ',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
lr = hyperparams.LogUniform(
default=1e-2,
lower=1e-05,
upper=1,
description='Maximum number of iterations. Default is 300 ',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
lr_decay = hyperparams.Hyperparameter[float](
default=0.95,
description='Maximum number of iterations. Default is 300 ',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
max_valid = hyperparams.Hyperparameter[int](
default=10,
description='Maximum number of iterations. Default is 300 ',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
valid_loss_weight = hyperparams.Hyperparameter[float](
default=0.5,
description='Maximum number of iterations. Default is 300 ',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
class RNNHyperparams(hyperparams.Hyperparams):
n_batch = hyperparams.Hyperparameter[int](
default=1,
description='Maximum number of batch size',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
n_max_epoch = hyperparams.Hyperparameter[int](
default=1000,
description='Maximum number of Epochs. Default is 300 ',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
n_max_epoch_total = hyperparams.Hyperparameter[int](
default=100,
description='Maximum number of total Epoches. Default is 300 ',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
n_neurons = hyperparams.Hyperparameter[int](
default=256,
description='Neurons in hidden layers',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
n_input_dim = hyperparams.Hyperparameter[int](
default=1,
description='Number of input dimension',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
n_dense_dim = hyperparams.Hyperparameter[int](
default=128,
description='Size of fully-connected layers',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
n_output_dim = hyperparams.Hyperparameter[int](
default=3,
description='output dimension',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
n_patience = hyperparams.Hyperparameter[int](
default=100,
description='Number of patience',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
n_lr_decay = hyperparams.Hyperparameter[int](
default=5,
description='number of Learning rate decay',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
lr = hyperparams.LogUniform(
default=1e-2,
lower=1e-05,
upper=1,
description='Learning rate',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
lr_decay = hyperparams.Hyperparameter[float](
default=0.95,
description='learning rate decay',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
max_valid = hyperparams.Hyperparameter[int](
default=10,
description='Maximum valid number of iterations. Default is 300 ',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
valid_loss_weight = hyperparams.Hyperparameter[float](
default=0.5,
description='Loss weight of validation set',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
])
class Hyperparams(hyperparams.Hyperparams):
hidden_layer_sizes = hyperparams.Hyperparameter[List](
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter',
],
default=List([30, 30]),
)
activation = hyperparams.Enumeration[str](
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter',
],
values=['identity', 'logistic', 'tanh', 'relu'],
default='relu',
description='Activation function for the hidden layer.')
solver = hyperparams.Enumeration[str](semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter',
],
values=['lbfgs', 'sgd', 'adam'],
default='adam',
description='')
learning_rate = hyperparams.Enumeration[str](
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter',
],
values=['constant', 'invscaling', 'adaptive'],
default='constant',
description='')
alpha = hyperparams.Hyperparameter[float](semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter',
],
default=0.0001,
description='')
beta_1 = hyperparams.Hyperparameter[float](semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter',
],
default=0.9,
description='')
beta_2 = hyperparams.Hyperparameter[float](semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter',
],
default=0.999,
description='')
learning_rate_init = hyperparams.Hyperparameter[float](
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter',
],
default=0.001,
)
tol = hyperparams.Hyperparameter[float](
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter',
],
default=0.0001,
)
max_iter = hyperparams.UniformInt(
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/ControlParameter',
'https://metadata.datadrivendiscovery.org/types/ResourcesUseParameter',
],
default=200,
lower=1,
upper=800,
description='The maximum number of passes over the training data ')
early_stopping = hyperparams.Hyperparameter[bool](
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/ControlParameter',
],
default=False,
)
shuffle = hyperparams.Hyperparameter[bool](
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/ControlParameter',
],
default=True,
description=
'Whether or not the training data should be shuffled after each iteration. '
)
warm_start = hyperparams.Hyperparameter[bool](
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/ControlParameter',
],
default=False,
description=
'When set to True, reuse the solution of the previous call to fit as initialization, otherwise, just erase the previous solution. '
)
#hidden_layer_sizes=hyperparams.Hyperparameter[Tuple[int, int]](
# default=(100,1)
#)
epsilon = hyperparams.LogUniform(
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter',
],
default=1e-8,
lower=1e-08,
upper=0.1,
description=
'Value for numerical stability in adam. Only used when solver=’adam’')
use_columns = hyperparams.Set(
elements=hyperparams.Hyperparameter[int](-1),
default=(),
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/ControlParameter'
],
description=
"A set of column indices to force primitive to operate on. If any specified column cannot be parsed, it is skipped.",
)
exclude_columns = hyperparams.Set(
elements=hyperparams.Hyperparameter[int](-1),
default=(),
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/ControlParameter'
],
description=
"A set of column indices to not operate on. Applicable only if \"use_columns\" is not provided.",
)
return_result = hyperparams.Enumeration(
values=['append', 'replace', 'new'],
default='replace',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/ControlParameter'
],
description=
"Should parsed columns be appended, should they replace original columns, or should only parsed columns be returned?",
)
use_semantic_types = hyperparams.UniformBool(
default=False,
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/ControlParameter'
],
description=
"Controls whether semantic_types metadata will be used for filtering columns in input dataframe."
)
add_index_columns = hyperparams.UniformBool(
default=True,
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/ControlParameter'
],
description=
"Also include primary index columns if input data has them. Applicable only if \"return_result\" is set to \"new\".",
)
lower=5,
upper=50,
default=30,
lower_inclusive=True,
upper_inclusive=True,
description='The maximum depth of the tree.',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
],
),
learning_rate=hyperparams.LogUniform(
lower=1e-4,
upper=1e-1,
default=0.05,
lower_inclusive=True,
upper_inclusive=True,
description=r'Boosting learning rate (xgb\`s \"eta\")',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
],
),
gamma=hyperparams.Constant[float](
default=0.0,
description=
'Minimum loss reduction required to make a further partition on a leaf node of the tree',
semantic_types=[
'https://metadata.datadrivendiscovery.org/types/TuningParameter'
],
),
min_child_weight=hyperparams.Constant[int](
default=1,
