com.eharmony.aloha.models.multilabel
Contains information about the labels to be used for predictions, and problems encountered while trying to get those labels.
Contains information about the labels to be used for predictions, and problems encountered while trying to get those labels.
type of label or class
indices into the sequence of all labels seen during training. These should be sorted in ascending order.
labels for which a prediction should be produced. labels are parallel to
indices so indices(i)
is the index associated with labels(i)
.
a sequence of labels derived from the input data that could not be found in the sequence of all labels seen during training.
any problems encountered when trying to get the labels. This should only be present when the caller indicates labels should be embedded in the input data passed to the prediction function in the MultilabelModel.
Combine the missing variables found into a set.
Combine the missing variables found into a set.
type of label or class
labels and information about the labels.
missing features from
a set of missing features
Get the labels and information about the labels.
Get the labels and information about the labels.
input type of the model
type of label or class
an input from which label information should be derived if labelsOfInterest is not empty.
an optional function used to extract label information from the input a
.
a mapping from label to index into the sequence of all labels seen during training.
label information related to all labels seen at training time. If
labelsOfInterest
is not provided, this information will be used.
labels and information about the labels.
Get labels from the input for which a prediction should be produced.
Get labels from the input for which a prediction should be produced. If
labelsOfInterest
produces a label not in the training set, it will not
be present in the prediction output but it will appear in
LabelsAndInfo.labelsNotInTrainingSet
.
input type of the model
type of label or class
the example provided to the model
a function used to extract labels for which a prediction should be produced.
mapping from Label to index into the sequence of all labels seen in the training set.
labels and information about the labels.
Report that no prediction attempt was made because of issues with the labels.
Report that no prediction attempt was made because of issues with the labels.
upper bound on model output type B
type of label or class
output type of the model.
An identifier for the model. Used in error reporting.
labels and information about the labels.
an auditor used to audit the output.
a SubValue indicating failure.
Report that a Throwable
was thrown while invoking the predictor
Report that a Throwable
was thrown while invoking the predictor
upper bound on model output type B
type of label or class
output type of the model.
An identifier for the model. Used in error reporting.
labels and information about the labels.
missing features from RegressionFeatures
the error the occurred in the predictor.
an auditor used to audit the output.
a SubValue indicating failure.
Report that the model succeeded.
Report that the model succeeded.
upper bound on model output type B
type of label or class
output type of the model.
An identifier for the model. Used in score reporting.
labels and information about the labels.
missing features from
the prediction(s) made by the embedded predictor.
an auditor used to audit the output.
a SubValue indicating success.
Report that a prediction could not be made because too many missing features were encountered.
Report that a prediction could not be made because too many missing features were encountered.
upper bound on model output type B
type of label or class
output type of the model.
An identifier for the model. Used in error reporting.
labels and information about the labels.
missing features from
an auditor used to audit the output.
a SubValue indicating failure.