Invokes the specified Machine Learning model that scores the input data.

To use this block, the Machine Learning application needs to be available with the respective Machine Learning models in the tenant.

If the Machine Learning model does not yet exist, use the Machine Learning application to add it. If you have added the Machine Learning model while your Analytics Builder model was still in edit mode, exit the model editor and then edit your Analytics Builder model once more. This refreshes the list of available Machine Learning models and you can then select the newly added model.

Block inputs correspond to the Machine Learning model's inputs (that are marked Active) in the order specified by the Machine Learning model. All inputs used by the model must be connected. Outputs correspond to the outputs as specified by the Machine Learning model. If a PMML output is specified as "JSON", then the block outputs a string version of the JSON, but the properties of the object are also available as extra values which can be extracted using the Extract Property block, which is the recommended way of unpacking multiple values from such an output. Currently, timestamp inputs are not supported.