def test_add_to_registry(self):
class Model1:
algorithm_short_name = ""
model1 = {
"task_name": "binary_classification",
"model_class": Model1,
"model_params": {},
"required_preprocessing": {},
"additional": {},
}
AlgorithmsRegistry.add(**model1)
"trees_in_step": 100,
"max_steps": 50,
"early_stopping_rounds": 50,
"max_rows_limit": None,
"max_cols_limit": None,
}
required_preprocessing = [
"missing_values_inputation",
"convert_categorical",
"target_as_integer",
]
AlgorithmsRegistry.add(
BINARY_CLASSIFICATION,
ExtraTreesAlgorithm,
et_params,
required_preprocessing,
additional,
classification_default_params,
)
AlgorithmsRegistry.add(
MULTICLASS_CLASSIFICATION,
ExtraTreesAlgorithm,
et_params,
required_preprocessing,
additional,
classification_default_params,
)
#
# REGRESSION
"max_cols_limit": None,
}
required_preprocessing = [
"missing_values_inputation",
"convert_categorical",
"datetime_transform",
"text_transform",
"scale",
"target_as_integer",
]
AlgorithmsRegistry.add(
BINARY_CLASSIFICATION,
NeuralNetworkAlgorithm,
nn_params,
required_preprocessing,
additional,
default_nn_params,
)
required_preprocessing = [
"missing_values_inputation",
"convert_categorical",
"datetime_transform",
"text_transform",
"scale",
"target_as_one_hot",
]
AlgorithmsRegistry.add(
MULTICLASS_CLASSIFICATION,
NeuralNetworkAlgorithm,
self.library_version = sklearn.__version__
self.max_iters = additional.get("max_steps", 1)
self.model = DummyRegressor(strategy="mean")
def file_extenstion(self):
return "baseline"
additional = {"max_steps": 1, "max_rows_limit": None, "max_cols_limit": None}
required_preprocessing = ["target_as_integer"]
AlgorithmsRegistry.add(
BINARY_CLASSIFICATION,
BaselineClassifierAlgorithm,
{},
required_preprocessing,
additional,
)
AlgorithmsRegistry.add(
MULTICLASS_CLASSIFICATION,
BaselineClassifierAlgorithm,
{},
required_preprocessing,
additional,
)
AlgorithmsRegistry.add(REGRESSION, BaselineRegressorAlgorithm, {}, {},
additional)
"early_stopping_rounds": 50,
"max_rows_limit": None,
"max_cols_limit": None,
}
required_preprocessing = [
"missing_values_inputation",
"convert_categorical",
"datetime_transform",
"text_transform",
"target_as_integer",
]
AlgorithmsRegistry.add(
BINARY_CLASSIFICATION,
XgbAlgorithm,
xgb_bin_class_params,
required_preprocessing,
additional,
classification_bin_default_params,
)
AlgorithmsRegistry.add(
MULTICLASS_CLASSIFICATION,
XgbAlgorithm,
xgb_multi_class_params,
required_preprocessing,
additional,
classification_multi_default_params,
)
regression_required_preprocessing = [
"missing_values_inputation",
"trees_in_step": 1,
"train_cant_improve_limit": 0,
"max_steps": 1,
"max_rows_limit": None,
"max_cols_limit": None,
}
required_preprocessing = [
"missing_values_inputation",
"convert_categorical",
"target_as_integer",
]
AlgorithmsRegistry.add(
BINARY_CLASSIFICATION,
DecisionTreeAlgorithm,
dt_params,
required_preprocessing,
additional,
classification_default_params,
)
AlgorithmsRegistry.add(
MULTICLASS_CLASSIFICATION,
DecisionTreeAlgorithm,
dt_params,
required_preprocessing,
additional,
classification_default_params,
)
dt_regression_params = {
"criterion": [
"convert_categorical",
"target_as_integer",
]
lgbm_multi_params = copy.deepcopy(lgbm_bin_params)
lgbm_multi_params["objective"] = ["multiclass"]
lgbm_multi_params["metric"] = ["multi_logloss", "multi_error"]
lgbr_params = copy.deepcopy(lgbm_bin_params)
lgbr_params["objective"] = ["regression"]
lgbr_params["metric"] = ["l1", "l2"]
AlgorithmsRegistry.add(
BINARY_CLASSIFICATION,
LightgbmAlgorithm,
lgbm_bin_params,
required_preprocessing,
additional,
)
AlgorithmsRegistry.add(
MULTICLASS_CLASSIFICATION,
LightgbmAlgorithm,
lgbm_multi_params,
required_preprocessing,
additional,
)
regression_required_preprocessing = [
"missing_values_inputation",
"convert_categorical",
"train_cant_improve_limit": 5,
"max_steps": 500,
"max_rows_limit": None,
"max_cols_limit": None,
}
required_preprocessing = [
"missing_values_inputation",
"convert_categorical",
"target_as_integer",
"target_scale",
]
AlgorithmsRegistry.add(
BINARY_CLASSIFICATION,
RandomForestAlgorithm,
rf_params,
required_preprocessing,
additional,
)
AlgorithmsRegistry.add(
MULTICLASS_CLASSIFICATION,
RandomForestAlgorithm,
rf_params,
required_preprocessing,
additional,
)
#
# REGRESSION
#
"max_rounds": 10000,
"early_stopping_rounds": 50,
"max_rows_limit": None,
"max_cols_limit": None,
}
required_preprocessing = [
"missing_values_inputation",
"datetime_transform",
"text_transform",
"target_as_integer",
]
AlgorithmsRegistry.add(
BINARY_CLASSIFICATION,
CatBoostAlgorithm,
classification_params,
required_preprocessing,
additional,
classification_default_params,
)
multiclass_classification_params = copy.deepcopy(classification_params)
multiclass_classification_params["loss_function"] = ["MultiClass"]
multiclass_classification_params["depth"] = [3, 4, 5, 6]
multiclass_classification_params["learning_rate"] = [0.1, 0.15, 0.2]
multiclass_classification_default_params = copy.deepcopy(
classification_default_params)
multiclass_classification_default_params["loss_function"] = "MultiClass"
multiclass_classification_default_params["depth"] = 5
multiclass_classification_default_params["learning_rate"] = 0.15
)
df.to_csv(
os.path.join(model_file_path, f"{learner_name}_coefs.csv"), index=False
)
additional = {"max_steps": 1, "max_rows_limit": None, "max_cols_limit": None}
required_preprocessing = [
"missing_values_inputation",
"convert_categorical",
"scale",
"target_as_integer",
]
AlgorithmsRegistry.add(
BINARY_CLASSIFICATION, LinearAlgorithm, {}, required_preprocessing, additional
)
AlgorithmsRegistry.add(
MULTICLASS_CLASSIFICATION, LinearAlgorithm, {}, required_preprocessing, additional
)
regression_required_preprocessing = [
"missing_values_inputation",
"convert_categorical",
"scale",
"target_scale",
]
AlgorithmsRegistry.add(
REGRESSION,
LinearRegressorAlgorithm,
additional = {
"one_step": 50,
"train_cant_improve_limit": 5,
"max_steps": 500,
"max_rows_limit": None,
"max_cols_limit": None,
}
required_preprocessing = [
"missing_values_inputation",
"convert_categorical",
"target_preprocessing",
]
"""
AlgorithmsRegistry.add(
BINARY_CLASSIFICATION,
CatBoostAlgorithm,
bin_class_params,
required_preprocessing,
additional,
)
"""
# switch off for now
# maybe my misuse or bug https://github.com/catboost/catboost/issues/861
"""
multi_class_params = copy.deepcopy(bin_class_params)
multi_class_params["loss_function"] = ["MultiClass"]
default_params = {"n_neighbors": 5, "weights": "uniform"}
additional = {"max_rows_limit": 100000, "max_cols_limit": 100}
required_preprocessing = [
"missing_values_inputation",
"convert_categorical",
"scale",
"target_as_integer",
]
AlgorithmsRegistry.add(
BINARY_CLASSIFICATION,
KNeighborsAlgorithm,
knn_params,
required_preprocessing,
additional,
default_params,
)
AlgorithmsRegistry.add(
MULTICLASS_CLASSIFICATION,
KNeighborsAlgorithm,
knn_params,
required_preprocessing,
additional,
default_params,
)
AlgorithmsRegistry.add(
REGRESSION,
KNeighborsRegressorAlgorithm,