You have executed a model node which has generated a model nugget connected to your stream's data source. What will allow the stream to score data with this new nugget?

Question

Options

• ARemove any partition node to allow all data to be scored.
• BAdd a Discriminant node to use the nugget model for scoring.
• CRemove the Reclassify node to use the nugget model for scoring.
• DAdd an output node downstream from the nugget.

• A
  7% (4)
• B
  3% (2)
• C
  74% (43)
• D
  16% (9)

Explanation

Removing the Reclassify node clears the path for the model nugget to score incoming data because the Reclassify node - typically used to transform the target field during model training - conflicts with the scoring process. When scoring new data, the target field is absent or in its original form, so a Reclassify node acting on that field causes the stream to fail or produce incorrect output; deleting it lets the nugget receive the raw predictor fields it expects and apply its learned rules directly.

Why the distractors are wrong:

• A (Remove partition node): Partition nodes split data for training/testing but don't block a nugget from scoring; removing one changes which records flow through but doesn't resolve a field-transformation conflict.
• B (Add a Discriminant node): A Discriminant node is a modeling node for discriminant analysis - adding a second modeler to a stream that already has a nugget is unnecessary and off-topic.
• D (Add an output node): While an output node is eventually needed to view or export results, the stream can already pass data through the nugget for scoring without one being the blocking factor described here.

Memory tip: Think "Reclassify = training-time only." If you reclassify a field to build a model, that same transformation becomes an obstacle when you switch the stream to score mode - so remove it to let the nugget work cleanly.

No community discussion yet for this question.