FMClassificationTrainingSummary#

class pyspark.ml.classification.FMClassificationTrainingSummary(java_obj=None)[source]#

Abstraction for FMClassifier Training results.

New in version 3.1.0.

Methods

`fMeasureByLabel`([beta])	Returns f-measure for each label (category).
`weightedFMeasure`([beta])	Returns weighted averaged f-measure.

Attributes

`accuracy`	Returns accuracy.
`areaUnderROC`	Computes the area under the receiver operating characteristic (ROC) curve.
`fMeasureByThreshold`	Returns a dataframe with two fields (threshold, F-Measure) curve with beta = 1.0.
`falsePositiveRateByLabel`	Returns false positive rate for each label (category).
`labelCol`	Field in "predictions" which gives the true label of each instance.
`labels`	Returns the sequence of labels in ascending order.
`objectiveHistory`	Objective function (scaled loss + regularization) at each iteration.
`pr`	Returns the precision-recall curve, which is a Dataframe containing two fields recall, precision with (0.0, 1.0) prepended to it.
`precisionByLabel`	Returns precision for each label (category).
`precisionByThreshold`	Returns a dataframe with two fields (threshold, precision) curve.
`predictionCol`	Field in "predictions" which gives the prediction of each class.
`predictions`	Dataframe outputted by the model's transform method.
`recallByLabel`	Returns recall for each label (category).
`recallByThreshold`	Returns a dataframe with two fields (threshold, recall) curve.
`roc`	Returns the receiver operating characteristic (ROC) curve, which is a Dataframe having two fields (FPR, TPR) with (0.0, 0.0) prepended and (1.0, 1.0) appended to it.
`scoreCol`	Field in "predictions" which gives the probability or raw prediction of each class as a vector.
`totalIterations`	Number of training iterations until termination.
`truePositiveRateByLabel`	Returns true positive rate for each label (category).
`weightCol`	Field in "predictions" which gives the weight of each instance as a vector.
`weightedFalsePositiveRate`	Returns weighted false positive rate.
`weightedPrecision`	Returns weighted averaged precision.
`weightedRecall`	Returns weighted averaged recall.
`weightedTruePositiveRate`	Returns weighted true positive rate.

Methods Documentation

fMeasureByLabel(beta=1.0)#: Returns f-measure for each label (category).

New in version 3.1.0.

weightedFMeasure(beta=1.0)#: Returns weighted averaged f-measure.

New in version 3.1.0.

Attributes Documentation

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

New in version 3.1.0.

areaUnderROC#: Computes the area under the receiver operating characteristic (ROC) curve.

New in version 3.1.0.

fMeasureByThreshold#: Returns a dataframe with two fields (threshold, F-Measure) curve with beta = 1.0.

New in version 3.1.0.

falsePositiveRateByLabel#: Returns false positive rate for each label (category).

New in version 3.1.0.

labelCol#: Field in “predictions” which gives the true label of each instance.

New in version 3.1.0.

labels#

Returns the sequence of labels in ascending order. This order matches the order used in metrics which are specified as arrays over labels, e.g., truePositiveRateByLabel.

New in version 3.1.0.

Notes

In most cases, it will be values {0.0, 1.0, …, numClasses-1}, However, if the training set is missing a label, then all of the arrays over labels (e.g., from truePositiveRateByLabel) will be of length numClasses-1 instead of the expected numClasses.

objectiveHistory#: Objective function (scaled loss + regularization) at each iteration. It contains one more element, the initial state, than number of iterations.

New in version 3.1.0.

pr#: Returns the precision-recall curve, which is a Dataframe containing two fields recall, precision with (0.0, 1.0) prepended to it.

New in version 3.1.0.

precisionByLabel#: Returns precision for each label (category).

New in version 3.1.0.

precisionByThreshold#: Returns a dataframe with two fields (threshold, precision) curve. Every possible probability obtained in transforming the dataset are used as thresholds used in calculating the precision.

New in version 3.1.0.

predictionCol#: Field in “predictions” which gives the prediction of each class.

New in version 3.1.0.

predictions#: Dataframe outputted by the model’s transform method.

New in version 3.1.0.

recallByLabel#: Returns recall for each label (category).

New in version 3.1.0.

recallByThreshold#: Returns a dataframe with two fields (threshold, recall) curve. Every possible probability obtained in transforming the dataset are used as thresholds used in calculating the recall.

New in version 3.1.0.

roc#

Returns the receiver operating characteristic (ROC) curve, which is a Dataframe having two fields (FPR, TPR) with (0.0, 0.0) prepended and (1.0, 1.0) appended to it.

New in version 3.1.0.

Notes

Wikipedia reference

scoreCol#: Field in “predictions” which gives the probability or raw prediction of each class as a vector.

New in version 3.1.0.

totalIterations#: Number of training iterations until termination.

New in version 3.1.0.

truePositiveRateByLabel#: Returns true positive rate for each label (category).

New in version 3.1.0.

weightCol#: Field in “predictions” which gives the weight of each instance as a vector.

New in version 3.1.0.

weightedFalsePositiveRate#: Returns weighted false positive rate.

New in version 3.1.0.

weightedPrecision#: Returns weighted averaged precision.

New in version 3.1.0.

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

New in version 3.1.0.

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

New in version 3.1.0.