MulticlassMetrics¶
-
class
pyspark.mllib.evaluation.
MulticlassMetrics
(predictionAndLabels)[source]¶ Evaluator for multiclass classification.
- Parameters
predictionAndLabels – an RDD of prediction, label, optional weight and optional probability.
>>> 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...
New in version 1.4.0.
Methods
Attributes
Methods Documentation
-
call
(name, *a)¶ Call method of java_model
-
confusionMatrix
()[source]¶ Returns confusion matrix: predicted classes are in columns, they are ordered by class label ascending, as in “labels”.
New in version 1.4.0.
-
falsePositiveRate
(label)[source]¶ Returns false positive rate for a given label (category).
New in version 1.4.0.
-
truePositiveRate
(label)[source]¶ Returns true positive rate for a given label (category).
New in version 1.4.0.
Attributes Documentation
-
accuracy
¶ Returns accuracy (equals to the total number of correctly classified instances out of the total number of instances).
New in version 2.0.0.
-
weightedFalsePositiveRate
¶ Returns weighted false positive rate.
New in version 1.4.0.
-
weightedPrecision
¶ Returns weighted averaged precision.
New in version 1.4.0.
-
weightedRecall
¶ Returns weighted averaged recall. (equals to precision, recall and f-measure)
New in version 1.4.0.
-
weightedTruePositiveRate
¶ Returns weighted true positive rate. (equals to precision, recall and f-measure)
New in version 1.4.0.