def predictWithQualityMetrics(self, trainingResult, testData,testGroundTruth):
# Retrieve predicted labels
predictedLabels = BinarySVM.predict (self, trainingResult, testData)
# Create a quality metric set object to compute quality metrics of the SVM algorithm
qualityMetricSet = quality_metric_set.Batch()
input = qualityMetricSet.getInputDataCollection().getInput(quality_metric_set.confusionMatrix)
input.set(binary_confusion_matrix.predictedLabels, predictedLabels)
input.set(binary_confusion_matrix.groundTruthLabels, testGroundTruth)
# Compute quality metrics and get the quality metrics
# returns ResultCollection class from svm.quality_metric_set
qualityMetricSetResult = qualityMetricSet.compute()
return predictedLabels, qualityMetricSetResult
def _computeTwoclassQualityMetrics(self):
# Alg object for quality metrics computation
quality_alg = twoclass_quality.Batch()
# Get access to the input parameter
input = quality_alg.getInputDataCollection().getInput(
twoclass_quality.confusionMatrix)
# Pass ground truth and predictions as input
input.set(binary_confusion_matrix.groundTruthLabels, self._truth)
input.set(binary_confusion_matrix.predictedLabels, self._predictions)
# Compute confusion matrix
confusion = quality_alg.compute().getResult(twoclass_quality.confusionMatrix)
# Retrieve quality metrics from the confusion matrix
metrics = confusion.get(binary_confusion_matrix.binaryMetrics)
# Convert the metrics into a Python namedtuple and return it
block = BlockDescriptor_Float64()
metrics.getBlockOfRows(0, 1, readOnly, block)
x = block.getArray().flatten()
self._metrics = TwoClassMetrics(*x)
metrics.releaseBlockOfRows(block)