ClassificationEvaluationMetrics

Metrics for classification evaluation results.

Fields
confidenceMetrics[] object (ConfidenceMetrics)

Metrics for each confidenceThreshold in 0.00,0.05,0.10,...,0.95,0.96,0.97,0.98,0.99 and positionThreshold = INT32_MAX_VALUE.

ROC and precision-recall curves, and other aggregated metrics are derived from them. The confidence metrics entries may also be supplied for additional values of positionThreshold, but from these no aggregated metrics are computed.

confusionMatrix object (ConfusionMatrix)

Confusion matrix of the evaluation.

auPrc number

The Area Under Precision-Recall Curve metric. Micro-averaged for the overall evaluation.

auRoc number

The Area Under Receiver Operating Characteristic curve metric. Micro-averaged for the overall evaluation.

logLoss number

The log Loss metric.

JSON representation 
{
  "confidenceMetrics": [
    {
      object (ConfidenceMetrics)
    }
  ],
  "confusionMatrix": {
    object (ConfusionMatrix)
  },
  "auPrc": number,
  "auRoc": number,
  "logLoss": number
}

ConfidenceMetrics

Fields
confusionMatrix object (ConfusionMatrix)

Confusion matrix of the evaluation for this confidenceThreshold.

confidenceThreshold number

Metrics are computed with an assumption that the Model never returns predictions with score lower than this value.

maxPredictions integer

Metrics are computed with an assumption that the Model always returns at most this many predictions (ordered by their score, descendingly), but they all still need to meet the confidenceThreshold.

recall number

Recall (True Positive Rate) for the given confidence threshold.

precision number

Precision for the given confidence threshold.

falsePositiveRate number

False Positive Rate for the given confidence threshold.

f1Score number

The harmonic mean of recall and precision. For summary metrics, it computes the micro-averaged F1 score.

f1ScoreMicro number

Micro-averaged F1 Score.

f1ScoreMacro number

Macro-averaged F1 Score.

recallAt1 number

The Recall (True Positive Rate) when only considering the label that has the highest prediction score and not below the confidence threshold for each DataItem.

precisionAt1 number

The precision when only considering the label that has the highest prediction score and not below the confidence threshold for each DataItem.

falsePositiveRateAt1 number

The False Positive Rate when only considering the label that has the highest prediction score and not below the confidence threshold for each DataItem.

f1ScoreAt1 number

The harmonic mean of recallAt1 and precisionAt1.

truePositiveCount string (int64 format)

The number of Model created labels that match a ground truth label.

falsePositiveCount string (int64 format)

The number of Model created labels that do not match a ground truth label.

falseNegativeCount string (int64 format)

The number of ground truth labels that are not matched by a Model created label.

trueNegativeCount string (int64 format)

The number of labels that were not created by the Model, but if they would, they would not match a ground truth label.

JSON representation 
{
  "confusionMatrix": {
    object (ConfusionMatrix)
  },
  "confidenceThreshold": number,
  "maxPredictions": integer,
  "recall": number,
  "precision": number,
  "falsePositiveRate": number,
  "f1Score": number,
  "f1ScoreMicro": number,
  "f1ScoreMacro": number,
  "recallAt1": number,
  "precisionAt1": number,
  "falsePositiveRateAt1": number,
  "f1ScoreAt1": number,
  "truePositiveCount": string,
  "falsePositiveCount": string,
  "falseNegativeCount": string,
  "trueNegativeCount": string
}