MulticlassMetrics¶

class pyspark.mllib.evaluation.MulticlassMetrics(predictionAndLabels: pyspark.rdd.RDD[Tuple[float, float]])¶

Evaluator for multiclass classification.

Parameters

predictionAndLabelspyspark.RDD: an RDD of prediction, label, optional weight and optional probability.

Examples

>>> predictionAndLabels = sc.parallelize([(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)])
>>> metrics = MulticlassMetrics(predictionAndLabels)
>>> metrics.confusionMatrix().toArray()
array([[ 2.,  1.,  1.],
       [ 1.,  3.,  0.],
       [ 0.,  0.,  1.]])
>>> metrics.falsePositiveRate(0.0)
0.2...
>>> metrics.precision(1.0)
0.75...
>>> metrics.recall(2.0)
1.0...
>>> metrics.fMeasure(0.0, 2.0)
0.52...
>>> metrics.accuracy
0.66...
>>> metrics.weightedFalsePositiveRate
0.19...
>>> metrics.weightedPrecision
0.68...
>>> metrics.weightedRecall
0.66...
>>> metrics.weightedFMeasure()
0.66...
>>> metrics.weightedFMeasure(2.0)
0.65...
>>> predAndLabelsWithOptWeight = sc.parallelize([(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)])
>>> metrics = MulticlassMetrics(predAndLabelsWithOptWeight)
>>> metrics.confusionMatrix().toArray()
array([[ 2.,  1.,  1.],
       [ 1.,  3.,  0.],
       [ 0.,  0.,  1.]])
>>> metrics.falsePositiveRate(0.0)
0.2...
>>> metrics.precision(1.0)
0.75...
>>> metrics.recall(2.0)
1.0...
>>> metrics.fMeasure(0.0, 2.0)
0.52...
>>> metrics.accuracy
0.66...
>>> metrics.weightedFalsePositiveRate
0.19...
>>> metrics.weightedPrecision
0.68...
>>> metrics.weightedRecall
0.66...
>>> metrics.weightedFMeasure()
0.66...
>>> metrics.weightedFMeasure(2.0)
0.65...
>>> predictionAndLabelsWithProbabilities = sc.parallelize([
...      (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])])
>>> metrics = MulticlassMetrics(predictionAndLabelsWithProbabilities)
>>> metrics.logLoss()
0.9682...

Methods

`call`(name, *a)	Call method of java_model
`confusionMatrix`()	Returns confusion matrix: predicted classes are in columns, they are ordered by class label ascending, as in “labels”.
`fMeasure`(label[, beta])	Returns f-measure.
`falsePositiveRate`(label)	Returns false positive rate for a given label (category).
`logLoss`([eps])	Returns weighted logLoss.
`precision`(label)	Returns precision.
`recall`(label)	Returns recall.
`truePositiveRate`(label)	Returns true positive rate for a given label (category).
`weightedFMeasure`([beta])	Returns weighted averaged f-measure.

Attributes

`accuracy`	Returns accuracy (equals to the total number of correctly classified instances out of the total number of instances).
`weightedFalsePositiveRate`	Returns weighted false positive rate.
`weightedPrecision`	Returns weighted averaged precision.
`weightedRecall`	Returns weighted averaged recall.
`weightedTruePositiveRate`	Returns weighted true positive rate.

Methods Documentation

call(name: str, *a: Any) → Any¶: Call method of java_model

confusionMatrix() → pyspark.mllib.linalg.Matrix ¶: Returns confusion matrix: predicted classes are in columns, they are ordered by class label ascending, as in “labels”.

fMeasure(label: float, beta: Optional[float] = None) → float¶: Returns f-measure.

falsePositiveRate(label: float) → float¶: Returns false positive rate for a given label (category).

logLoss(eps: float = 1e-15) → float¶: Returns weighted logLoss.

precision(label: float) → float¶: Returns precision.

recall(label: float) → float¶: Returns recall.

truePositiveRate(label: float) → float¶: Returns true positive rate for a given label (category).

weightedFMeasure(beta: Optional[float] = None) → float¶: Returns weighted averaged f-measure.

Attributes Documentation

accuracy¶: Returns accuracy (equals to the total number of correctly classified instances out of the total number of instances).

weightedFalsePositiveRate¶: Returns weighted false positive rate.

weightedPrecision¶: Returns weighted averaged precision.

weightedRecall¶: Returns weighted averaged recall. (equals to precision, recall and f-measure)

weightedTruePositiveRate¶: Returns weighted true positive rate. (equals to precision, recall and f-measure)

RegressionMetrics

RankingMetrics