You need to score future data using an association model. Which statement describes the data format requirement to do this?

Question

Options

• AAn association model built using transactional data can be used to score tabular data.
• BAn association model built using tabular data can be used to score transactional data.
• CAny association model can be used to score both transactional and tabular data: regardless of
• DAn association model built using transactional data can only be used to score transactional

• A
  11% (6)
• B
  80% (44)
• C
  4% (2)
• D
  5% (3)

Explanation

Option B is correct because in association modeling, the training data format determines scoring flexibility: a model trained on tabular data (rows and columns) is compatible with transactional data (item lists/baskets) at scoring time, giving it cross-format scoring capability.

Why the distractors are wrong:

• A reverses the relationship - a model built on transactional data cannot score tabular data; the compatibility only flows in the tabular-to-transactional direction, not the other way around.
• C overstates flexibility - it is false that any association model can score both formats; only tabular-trained models have this cross-format capability.
• D describes transactional-trained models as scoring only transactional data, which is actually true but does not answer the question about which model can score future transactional data when you have a choice - making it a misleading but incomplete answer compared to B.

Memory tip: Think "Tabular is the Transportable one" - a tabular-built association model can travel to score transactional data, but a transactional-built model is stuck in its own lane.

No community discussion yet for this question.