# Copyright (C) Microsoft Corporation. All rights reserved.
# Licensed under the MIT License. See LICENSE in project root for information.
import sys
if sys.version >= '3':
basestring = str
from pyspark.ml.param.shared import *
from pyspark import keyword_only
from pyspark.ml.util import JavaMLReadable, JavaMLWritable
from pyspark.ml.wrapper import JavaTransformer, JavaEstimator, JavaModel
from pyspark.ml.common import inherit_doc
from mmlspark.Utils import *
from mmlspark.TypeConversionUtils import generateTypeConverter, complexTypeConverter
[docs]@inherit_doc
class TrainRegressor(ComplexParamsMixin, JavaMLReadable, JavaMLWritable, JavaEstimator):
"""
Use ``TrainRegressor`` to train a regression model on a dataset.
Below is an example that uses ``TrainRegressor``. Given a DataFrame,
myDataFrame, with a label column, "MyLabel", split the DataFrame into
train and test sets. Train a regressor on the dataset with a solver,
such as l-bfgs:
>>> from mmlspark.TrainRegressor import TrainRegressor
>>> from pysppark.ml.regression import LinearRegression
>>> lr = LinearRegression().setSolver("l-bfgs").setRegParam(0.1).setElasticNetParam(0.3)
>>> model = TrainRegressor(model=lr, labelCol="MyLabel", numFeatures=1 << 18).fit(train)
Now that you have a model, you can score the regressor on the test data:
>>> scoredData = model.transform(test)
Args:
labelCol (str): The name of the label column
model (object): Regressor to run
numFeatures (int): Number of features to hash to (default: 0)
"""
@keyword_only
def __init__(self, labelCol=None, model=None, numFeatures=0):
super(TrainRegressor, self).__init__()
self._java_obj = self._new_java_obj("com.microsoft.ml.spark.TrainRegressor")
self._cache = {}
self.labelCol = Param(self, "labelCol", "labelCol: The name of the label column")
self.model = Param(self, "model", "model: Regressor to run", generateTypeConverter("model", self._cache, complexTypeConverter))
self.numFeatures = Param(self, "numFeatures", "numFeatures: Number of features to hash to (default: 0)")
self._setDefault(numFeatures=0)
if hasattr(self, "_input_kwargs"):
kwargs = self._input_kwargs
else:
kwargs = self.__init__._input_kwargs
self.setParams(**kwargs)
[docs] @keyword_only
def setParams(self, labelCol=None, model=None, numFeatures=0):
"""
Set the (keyword only) parameters
Args:
labelCol (str): The name of the label column
model (object): Regressor to run
numFeatures (int): Number of features to hash to (default: 0)
"""
if hasattr(self, "_input_kwargs"):
kwargs = self._input_kwargs
else:
kwargs = self.__init__._input_kwargs
return self._set(**kwargs)
[docs] def setLabelCol(self, value):
"""
Args:
labelCol (str): The name of the label column
"""
self._set(labelCol=value)
return self
[docs] def getLabelCol(self):
"""
Returns:
str: The name of the label column
"""
return self.getOrDefault(self.labelCol)
[docs] def setModel(self, value):
"""
Args:
model (object): Regressor to run
"""
self._set(model=value)
return self
[docs] def getModel(self):
"""
Returns:
object: Regressor to run
"""
return self._cache.get("model", None)
[docs] def setNumFeatures(self, value):
"""
Args:
numFeatures (int): Number of features to hash to (default: 0)
"""
self._set(numFeatures=value)
return self
[docs] def getNumFeatures(self):
"""
Returns:
int: Number of features to hash to (default: 0)
"""
return self.getOrDefault(self.numFeatures)
[docs] @classmethod
def read(cls):
""" Returns an MLReader instance for this class. """
return JavaMMLReader(cls)
[docs] @staticmethod
def getJavaPackage():
""" Returns package name String. """
return "com.microsoft.ml.spark.TrainRegressor"
@staticmethod
def _from_java(java_stage):
module_name=TrainRegressor.__module__
module_name=module_name.rsplit(".", 1)[0] + ".TrainRegressor"
return from_java(java_stage, module_name)
def _create_model(self, java_model):
return TrainedRegressorModel(java_model)
[docs]class TrainedRegressorModel(ComplexParamsMixin, JavaModel, JavaMLWritable, JavaMLReadable):
"""
Model fitted by :class:`TrainRegressor`.
This class is left empty on purpose.
All necessary methods are exposed through inheritance.
"""
[docs] @classmethod
def read(cls):
""" Returns an MLReader instance for this class. """
return JavaMMLReader(cls)
[docs] @staticmethod
def getJavaPackage():
""" Returns package name String. """
return "com.microsoft.ml.spark.TrainedRegressorModel"
@staticmethod
def _from_java(java_stage):
module_name=TrainedRegressorModel.__module__
module_name=module_name.rsplit(".", 1)[0] + ".TrainedRegressorModel"
return from_java(java_stage, module_name)