Coverage for pyspark/ml/evaluation.py : 81%
Hot-keys on this page
r m x p toggle line displays
j k next/prev highlighted chunk
0 (zero) top of page
1 (one) first highlighted chunk
|
# # Licensed to the Apache Software Foundation (ASF) under one or more # contributor license agreements. See the NOTICE file distributed with # this work for additional information regarding copyright ownership. # The ASF licenses this file to You under the Apache License, Version 2.0 # (the "License"); you may not use this file except in compliance with # the License. You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. #
HasRawPredictionCol, HasFeaturesCol, HasWeightCol
'MulticlassClassificationEvaluator', 'MultilabelClassificationEvaluator', 'ClusteringEvaluator', 'RankingEvaluator']
""" Base class for evaluators that compute metrics from predictions.
.. versionadded:: 1.4.0 """ pass
def _evaluate(self, dataset): """ Evaluates the output.
Parameters ---------- dataset : :py:class:`pyspark.sql.DataFrame` a dataset that contains labels/observations and predictions
Returns ------- float metric """ raise NotImplementedError()
""" Evaluates the output with optional parameters.
.. versionadded:: 1.4.0
Parameters ---------- dataset : :py:class:`pyspark.sql.DataFrame` a dataset that contains labels/observations and predictions params : dict, optional an optional param map that overrides embedded params
Returns ------- float metric """ else: else:
def isLargerBetter(self): """ Indicates whether the metric returned by :py:meth:`evaluate` should be maximized (True, default) or minimized (False). A given evaluator may support multiple metrics which may be maximized or minimized. """
""" Base class for :py:class:`Evaluator`s that wrap Java/Scala implementations. """
""" Evaluates the output.
Parameters ---------- dataset : :py:class:`pyspark.sql.DataFrame` a dataset that contains labels/observations and predictions
Returns ------- float evaluation metric """
JavaMLReadable, JavaMLWritable): """ Evaluator for binary classification, which expects input columns rawPrediction, label and an optional weight column. The rawPrediction column can be of type double (binary 0/1 prediction, or probability of label 1) or of type vector (length-2 vector of raw predictions, scores, or label probabilities).
.. versionadded:: 1.4.0
Examples -------- >>> from pyspark.ml.linalg import Vectors >>> scoreAndLabels = map(lambda x: (Vectors.dense([1.0 - x[0], x[0]]), x[1]), ... [(0.1, 0.0), (0.1, 1.0), (0.4, 0.0), (0.6, 0.0), (0.6, 1.0), (0.6, 1.0), (0.8, 1.0)]) >>> dataset = spark.createDataFrame(scoreAndLabels, ["raw", "label"]) ... >>> evaluator = BinaryClassificationEvaluator() >>> evaluator.setRawPredictionCol("raw") BinaryClassificationEvaluator... >>> evaluator.evaluate(dataset) 0.70... >>> evaluator.evaluate(dataset, {evaluator.metricName: "areaUnderPR"}) 0.83... >>> bce_path = temp_path + "/bce" >>> evaluator.save(bce_path) >>> evaluator2 = BinaryClassificationEvaluator.load(bce_path) >>> str(evaluator2.getRawPredictionCol()) 'raw' >>> scoreAndLabelsAndWeight = map(lambda x: (Vectors.dense([1.0 - x[0], x[0]]), x[1], x[2]), ... [(0.1, 0.0, 1.0), (0.1, 1.0, 0.9), (0.4, 0.0, 0.7), (0.6, 0.0, 0.9), ... (0.6, 1.0, 1.0), (0.6, 1.0, 0.3), (0.8, 1.0, 1.0)]) >>> dataset = spark.createDataFrame(scoreAndLabelsAndWeight, ["raw", "label", "weight"]) ... >>> evaluator = BinaryClassificationEvaluator(rawPredictionCol="raw", weightCol="weight") >>> evaluator.evaluate(dataset) 0.70... >>> evaluator.evaluate(dataset, {evaluator.metricName: "areaUnderPR"}) 0.82... >>> evaluator.getNumBins() 1000 """
"metric name in evaluation (areaUnderROC|areaUnderPR)", typeConverter=TypeConverters.toString)
"(ROC curve, PR curve) in area computation. If 0, no down-sampling will " "occur. Must be >= 0.", typeConverter=TypeConverters.toInt)
metricName="areaUnderROC", weightCol=None, numBins=1000): """ __init__(self, \\*, rawPredictionCol="rawPrediction", labelCol="label", \ metricName="areaUnderROC", weightCol=None, numBins=1000) """ "org.apache.spark.ml.evaluation.BinaryClassificationEvaluator", self.uid)
def setMetricName(self, value): """ Sets the value of :py:attr:`metricName`. """ return self._set(metricName=value)
def getMetricName(self): """ Gets the value of metricName or its default value. """ return self.getOrDefault(self.metricName)
def setNumBins(self, value): """ Sets the value of :py:attr:`numBins`. """ return self._set(numBins=value)
def getNumBins(self): """ Gets the value of numBins or its default value. """
""" Sets the value of :py:attr:`labelCol`. """ return self._set(labelCol=value)
""" Sets the value of :py:attr:`rawPredictionCol`. """
def setWeightCol(self, value): """ Sets the value of :py:attr:`weightCol`. """ return self._set(weightCol=value)
metricName="areaUnderROC", weightCol=None, numBins=1000): """ setParams(self, \\*, rawPredictionCol="rawPrediction", labelCol="label", \ metricName="areaUnderROC", weightCol=None, numBins=1000) Sets params for binary classification evaluator. """ kwargs = self._input_kwargs return self._set(**kwargs)
JavaMLReadable, JavaMLWritable): """ Evaluator for Regression, which expects input columns prediction, label and an optional weight column.
.. versionadded:: 1.4.0
Examples -------- >>> scoreAndLabels = [(-28.98343821, -27.0), (20.21491975, 21.5), ... (-25.98418959, -22.0), (30.69731842, 33.0), (74.69283752, 71.0)] >>> dataset = spark.createDataFrame(scoreAndLabels, ["raw", "label"]) ... >>> evaluator = RegressionEvaluator() >>> evaluator.setPredictionCol("raw") RegressionEvaluator... >>> evaluator.evaluate(dataset) 2.842... >>> evaluator.evaluate(dataset, {evaluator.metricName: "r2"}) 0.993... >>> evaluator.evaluate(dataset, {evaluator.metricName: "mae"}) 2.649... >>> re_path = temp_path + "/re" >>> evaluator.save(re_path) >>> evaluator2 = RegressionEvaluator.load(re_path) >>> str(evaluator2.getPredictionCol()) 'raw' >>> scoreAndLabelsAndWeight = [(-28.98343821, -27.0, 1.0), (20.21491975, 21.5, 0.8), ... (-25.98418959, -22.0, 1.0), (30.69731842, 33.0, 0.6), (74.69283752, 71.0, 0.2)] >>> dataset = spark.createDataFrame(scoreAndLabelsAndWeight, ["raw", "label", "weight"]) ... >>> evaluator = RegressionEvaluator(predictionCol="raw", weightCol="weight") >>> evaluator.evaluate(dataset) 2.740... >>> evaluator.getThroughOrigin() False """ """metric name in evaluation - one of: rmse - root mean squared error (default) mse - mean squared error r2 - r^2 metric mae - mean absolute error var - explained variance.""", typeConverter=TypeConverters.toString)
"whether the regression is through the origin.", typeConverter=TypeConverters.toBoolean)
metricName="rmse", weightCol=None, throughOrigin=False): """ __init__(self, \\*, predictionCol="prediction", labelCol="label", \ metricName="rmse", weightCol=None, throughOrigin=False) """ "org.apache.spark.ml.evaluation.RegressionEvaluator", self.uid)
def setMetricName(self, value): """ Sets the value of :py:attr:`metricName`. """ return self._set(metricName=value)
def getMetricName(self): """ Gets the value of metricName or its default value. """ return self.getOrDefault(self.metricName)
def setThroughOrigin(self, value): """ Sets the value of :py:attr:`throughOrigin`. """ return self._set(throughOrigin=value)
def getThroughOrigin(self): """ Gets the value of throughOrigin or its default value. """
""" Sets the value of :py:attr:`labelCol`. """ return self._set(labelCol=value)
""" Sets the value of :py:attr:`predictionCol`. """
def setWeightCol(self, value): """ Sets the value of :py:attr:`weightCol`. """ return self._set(weightCol=value)
metricName="rmse", weightCol=None, throughOrigin=False): """ setParams(self, \\*, predictionCol="prediction", labelCol="label", \ metricName="rmse", weightCol=None, throughOrigin=False) Sets params for regression evaluator. """ kwargs = self._input_kwargs return self._set(**kwargs)
HasProbabilityCol, JavaMLReadable, JavaMLWritable): """ Evaluator for Multiclass Classification, which expects input columns: prediction, label, weight (optional) and probabilityCol (only for logLoss).
.. versionadded:: 1.5.0
Examples -------- >>> scoreAndLabels = [(0.0, 0.0), (0.0, 1.0), (0.0, 0.0), ... (1.0, 0.0), (1.0, 1.0), (1.0, 1.0), (1.0, 1.0), (2.0, 2.0), (2.0, 0.0)] >>> dataset = spark.createDataFrame(scoreAndLabels, ["prediction", "label"]) >>> evaluator = MulticlassClassificationEvaluator() >>> evaluator.setPredictionCol("prediction") MulticlassClassificationEvaluator... >>> evaluator.evaluate(dataset) 0.66... >>> evaluator.evaluate(dataset, {evaluator.metricName: "accuracy"}) 0.66... >>> evaluator.evaluate(dataset, {evaluator.metricName: "truePositiveRateByLabel", ... evaluator.metricLabel: 1.0}) 0.75... >>> evaluator.setMetricName("hammingLoss") MulticlassClassificationEvaluator... >>> evaluator.evaluate(dataset) 0.33... >>> mce_path = temp_path + "/mce" >>> evaluator.save(mce_path) >>> evaluator2 = MulticlassClassificationEvaluator.load(mce_path) >>> str(evaluator2.getPredictionCol()) 'prediction' >>> scoreAndLabelsAndWeight = [(0.0, 0.0, 1.0), (0.0, 1.0, 1.0), (0.0, 0.0, 1.0), ... (1.0, 0.0, 1.0), (1.0, 1.0, 1.0), (1.0, 1.0, 1.0), (1.0, 1.0, 1.0), ... (2.0, 2.0, 1.0), (2.0, 0.0, 1.0)] >>> dataset = spark.createDataFrame(scoreAndLabelsAndWeight, ["prediction", "label", "weight"]) >>> evaluator = MulticlassClassificationEvaluator(predictionCol="prediction", ... weightCol="weight") >>> evaluator.evaluate(dataset) 0.66... >>> evaluator.evaluate(dataset, {evaluator.metricName: "accuracy"}) 0.66... >>> predictionAndLabelsWithProbabilities = [ ... (1.0, 1.0, 1.0, [0.1, 0.8, 0.1]), (0.0, 2.0, 1.0, [0.9, 0.05, 0.05]), ... (0.0, 0.0, 1.0, [0.8, 0.2, 0.0]), (1.0, 1.0, 1.0, [0.3, 0.65, 0.05])] >>> dataset = spark.createDataFrame(predictionAndLabelsWithProbabilities, ["prediction", ... "label", "weight", "probability"]) >>> evaluator = MulticlassClassificationEvaluator(predictionCol="prediction", ... probabilityCol="probability") >>> evaluator.setMetricName("logLoss") MulticlassClassificationEvaluator... >>> evaluator.evaluate(dataset) 0.9682... """ "metric name in evaluation " "(f1|accuracy|weightedPrecision|weightedRecall|weightedTruePositiveRate| " "weightedFalsePositiveRate|weightedFMeasure|truePositiveRateByLabel| " "falsePositiveRateByLabel|precisionByLabel|recallByLabel|fMeasureByLabel| " "logLoss|hammingLoss)", typeConverter=TypeConverters.toString) "The class whose metric will be computed in truePositiveRateByLabel|" "falsePositiveRateByLabel|precisionByLabel|recallByLabel|fMeasureByLabel." " Must be >= 0. The default value is 0.", typeConverter=TypeConverters.toFloat) "The beta value used in weightedFMeasure|fMeasureByLabel." " Must be > 0. The default value is 1.", typeConverter=TypeConverters.toFloat) "log-loss is undefined for p=0 or p=1, so probabilities are clipped to " "max(eps, min(1 - eps, p)). " "Must be in range (0, 0.5). The default value is 1e-15.", typeConverter=TypeConverters.toFloat)
metricName="f1", weightCol=None, metricLabel=0.0, beta=1.0, probabilityCol="probability", eps=1e-15): """ __init__(self, \\*, predictionCol="prediction", labelCol="label", \ metricName="f1", weightCol=None, metricLabel=0.0, beta=1.0, \ probabilityCol="probability", eps=1e-15) """ "org.apache.spark.ml.evaluation.MulticlassClassificationEvaluator", self.uid)
def setMetricName(self, value): """ Sets the value of :py:attr:`metricName`. """
def getMetricName(self): """ Gets the value of metricName or its default value. """ return self.getOrDefault(self.metricName)
def setMetricLabel(self, value): """ Sets the value of :py:attr:`metricLabel`. """ return self._set(metricLabel=value)
def getMetricLabel(self): """ Gets the value of metricLabel or its default value. """ return self.getOrDefault(self.metricLabel)
def setBeta(self, value): """ Sets the value of :py:attr:`beta`. """ return self._set(beta=value)
def getBeta(self): """ Gets the value of beta or its default value. """ return self.getOrDefault(self.beta)
def setEps(self, value): """ Sets the value of :py:attr:`eps`. """ return self._set(eps=value)
def getEps(self): """ Gets the value of eps or its default value. """ return self.getOrDefault(self.eps)
""" Sets the value of :py:attr:`labelCol`. """ return self._set(labelCol=value)
""" Sets the value of :py:attr:`predictionCol`. """
def setProbabilityCol(self, value): """ Sets the value of :py:attr:`probabilityCol`. """ return self._set(probabilityCol=value)
def setWeightCol(self, value): """ Sets the value of :py:attr:`weightCol`. """ return self._set(weightCol=value)
metricName="f1", weightCol=None, metricLabel=0.0, beta=1.0, probabilityCol="probability", eps=1e-15): """ setParams(self, \\*, predictionCol="prediction", labelCol="label", \ metricName="f1", weightCol=None, metricLabel=0.0, beta=1.0, \ probabilityCol="probability", eps=1e-15) Sets params for multiclass classification evaluator. """ kwargs = self._input_kwargs return self._set(**kwargs)
JavaMLReadable, JavaMLWritable): """ Evaluator for Multilabel Classification, which expects two input columns: prediction and label.
.. versionadded:: 3.0.0
Notes ----- Experimental
Examples -------- >>> scoreAndLabels = [([0.0, 1.0], [0.0, 2.0]), ([0.0, 2.0], [0.0, 1.0]), ... ([], [0.0]), ([2.0], [2.0]), ([2.0, 0.0], [2.0, 0.0]), ... ([0.0, 1.0, 2.0], [0.0, 1.0]), ([1.0], [1.0, 2.0])] >>> dataset = spark.createDataFrame(scoreAndLabels, ["prediction", "label"]) ... >>> evaluator = MultilabelClassificationEvaluator() >>> evaluator.setPredictionCol("prediction") MultilabelClassificationEvaluator... >>> evaluator.evaluate(dataset) 0.63... >>> evaluator.evaluate(dataset, {evaluator.metricName: "accuracy"}) 0.54... >>> mlce_path = temp_path + "/mlce" >>> evaluator.save(mlce_path) >>> evaluator2 = MultilabelClassificationEvaluator.load(mlce_path) >>> str(evaluator2.getPredictionCol()) 'prediction' """ "metric name in evaluation " "(subsetAccuracy|accuracy|hammingLoss|precision|recall|f1Measure|" "precisionByLabel|recallByLabel|f1MeasureByLabel|microPrecision|" "microRecall|microF1Measure)", typeConverter=TypeConverters.toString) "The class whose metric will be computed in precisionByLabel|" "recallByLabel|f1MeasureByLabel. " "Must be >= 0. The default value is 0.", typeConverter=TypeConverters.toFloat)
metricName="f1Measure", metricLabel=0.0): """ __init__(self, \\*, predictionCol="prediction", labelCol="label", \ metricName="f1Measure", metricLabel=0.0) """ "org.apache.spark.ml.evaluation.MultilabelClassificationEvaluator", self.uid)
def setMetricName(self, value): """ Sets the value of :py:attr:`metricName`. """ return self._set(metricName=value)
def getMetricName(self): """ Gets the value of metricName or its default value. """ return self.getOrDefault(self.metricName)
def setMetricLabel(self, value): """ Sets the value of :py:attr:`metricLabel`. """ return self._set(metricLabel=value)
def getMetricLabel(self): """ Gets the value of metricLabel or its default value. """ return self.getOrDefault(self.metricLabel)
def setLabelCol(self, value): """ Sets the value of :py:attr:`labelCol`. """ return self._set(labelCol=value)
def setPredictionCol(self, value): """ Sets the value of :py:attr:`predictionCol`. """
metricName="f1Measure", metricLabel=0.0): """ setParams(self, \\*, predictionCol="prediction", labelCol="label", \ metricName="f1Measure", metricLabel=0.0) Sets params for multilabel classification evaluator. """ kwargs = self._input_kwargs return self._set(**kwargs)
JavaMLReadable, JavaMLWritable): """ Evaluator for Clustering results, which expects two input columns: prediction and features. The metric computes the Silhouette measure using the squared Euclidean distance.
The Silhouette is a measure for the validation of the consistency within clusters. It ranges between 1 and -1, where a value close to 1 means that the points in a cluster are close to the other points in the same cluster and far from the points of the other clusters.
.. versionadded:: 2.3.0
Examples -------- >>> from pyspark.ml.linalg import Vectors >>> featureAndPredictions = map(lambda x: (Vectors.dense(x[0]), x[1]), ... [([0.0, 0.5], 0.0), ([0.5, 0.0], 0.0), ([10.0, 11.0], 1.0), ... ([10.5, 11.5], 1.0), ([1.0, 1.0], 0.0), ([8.0, 6.0], 1.0)]) >>> dataset = spark.createDataFrame(featureAndPredictions, ["features", "prediction"]) ... >>> evaluator = ClusteringEvaluator() >>> evaluator.setPredictionCol("prediction") ClusteringEvaluator... >>> evaluator.evaluate(dataset) 0.9079... >>> featureAndPredictionsWithWeight = map(lambda x: (Vectors.dense(x[0]), x[1], x[2]), ... [([0.0, 0.5], 0.0, 2.5), ([0.5, 0.0], 0.0, 2.5), ([10.0, 11.0], 1.0, 2.5), ... ([10.5, 11.5], 1.0, 2.5), ([1.0, 1.0], 0.0, 2.5), ([8.0, 6.0], 1.0, 2.5)]) >>> dataset = spark.createDataFrame( ... featureAndPredictionsWithWeight, ["features", "prediction", "weight"]) >>> evaluator = ClusteringEvaluator() >>> evaluator.setPredictionCol("prediction") ClusteringEvaluator... >>> evaluator.setWeightCol("weight") ClusteringEvaluator... >>> evaluator.evaluate(dataset) 0.9079... >>> ce_path = temp_path + "/ce" >>> evaluator.save(ce_path) >>> evaluator2 = ClusteringEvaluator.load(ce_path) >>> str(evaluator2.getPredictionCol()) 'prediction' """ "metric name in evaluation (silhouette)", typeConverter=TypeConverters.toString) "Supported options: 'squaredEuclidean' and 'cosine'.", typeConverter=TypeConverters.toString)
metricName="silhouette", distanceMeasure="squaredEuclidean", weightCol=None): """ __init__(self, \\*, predictionCol="prediction", featuresCol="features", \ metricName="silhouette", distanceMeasure="squaredEuclidean", weightCol=None) """ "org.apache.spark.ml.evaluation.ClusteringEvaluator", self.uid)
metricName="silhouette", distanceMeasure="squaredEuclidean", weightCol=None): """ setParams(self, \\*, predictionCol="prediction", featuresCol="features", \ metricName="silhouette", distanceMeasure="squaredEuclidean", weightCol=None) Sets params for clustering evaluator. """ kwargs = self._input_kwargs return self._set(**kwargs)
def setMetricName(self, value): """ Sets the value of :py:attr:`metricName`. """ return self._set(metricName=value)
def getMetricName(self): """ Gets the value of metricName or its default value. """ return self.getOrDefault(self.metricName)
def setDistanceMeasure(self, value): """ Sets the value of :py:attr:`distanceMeasure`. """ return self._set(distanceMeasure=value)
def getDistanceMeasure(self): """ Gets the value of `distanceMeasure` """
""" Sets the value of :py:attr:`featuresCol`. """ return self._set(featuresCol=value)
""" Sets the value of :py:attr:`predictionCol`. """
def setWeightCol(self, value): """ Sets the value of :py:attr:`weightCol`. """
JavaMLReadable, JavaMLWritable): """ Evaluator for Ranking, which expects two input columns: prediction and label.
.. versionadded:: 3.0.0
Notes ----- Experimental
Examples -------- >>> scoreAndLabels = [([1.0, 6.0, 2.0, 7.0, 8.0, 3.0, 9.0, 10.0, 4.0, 5.0], ... [1.0, 2.0, 3.0, 4.0, 5.0]), ... ([4.0, 1.0, 5.0, 6.0, 2.0, 7.0, 3.0, 8.0, 9.0, 10.0], [1.0, 2.0, 3.0]), ... ([1.0, 2.0, 3.0, 4.0, 5.0], [])] >>> dataset = spark.createDataFrame(scoreAndLabels, ["prediction", "label"]) ... >>> evaluator = RankingEvaluator() >>> evaluator.setPredictionCol("prediction") RankingEvaluator... >>> evaluator.evaluate(dataset) 0.35... >>> evaluator.evaluate(dataset, {evaluator.metricName: "precisionAtK", evaluator.k: 2}) 0.33... >>> ranke_path = temp_path + "/ranke" >>> evaluator.save(ranke_path) >>> evaluator2 = RankingEvaluator.load(ranke_path) >>> str(evaluator2.getPredictionCol()) 'prediction' """ "metric name in evaluation " "(meanAveragePrecision|meanAveragePrecisionAtK|" "precisionAtK|ndcgAtK|recallAtK)", typeConverter=TypeConverters.toString) "The ranking position value used in meanAveragePrecisionAtK|precisionAtK|" "ndcgAtK|recallAtK. Must be > 0. The default value is 10.", typeConverter=TypeConverters.toInt)
metricName="meanAveragePrecision", k=10): """ __init__(self, \\*, predictionCol="prediction", labelCol="label", \ metricName="meanAveragePrecision", k=10) """ "org.apache.spark.ml.evaluation.RankingEvaluator", self.uid)
def setMetricName(self, value): """ Sets the value of :py:attr:`metricName`. """ return self._set(metricName=value)
def getMetricName(self): """ Gets the value of metricName or its default value. """ return self.getOrDefault(self.metricName)
def setK(self, value): """ Sets the value of :py:attr:`k`. """ return self._set(k=value)
def getK(self): """ Gets the value of k or its default value. """ return self.getOrDefault(self.k)
def setLabelCol(self, value): """ Sets the value of :py:attr:`labelCol`. """ return self._set(labelCol=value)
def setPredictionCol(self, value): """ Sets the value of :py:attr:`predictionCol`. """
metricName="meanAveragePrecision", k=10): """ setParams(self, \\*, predictionCol="prediction", labelCol="label", \ metricName="meanAveragePrecision", k=10) Sets params for ranking evaluator. """ kwargs = self._input_kwargs return self._set(**kwargs)
# The small batch size here ensures that we see multiple batches, # even in these small test examples: .master("local[2]")\ .appName("ml.evaluation tests")\ .getOrCreate() finally: except OSError: pass sys.exit(-1) |