Building a training quest

A training quest flow contains multiple tasks to prepare data, train, score, and evaluate the model.

To build the flow of the training quest, follow these steps:

  1. Import a previously loaded dataset.
  2. Transform data into a useful structure with preprocessing activities, such as normalization, transformation, or reduction.
  3. Explore the data by visualizing it into various plot types.
  4. Apply an algorithm and tune hyperparameters.
  5. Apply a train model activity.
  6. Score and evaluate the model.
  7. Compare models if multiple models were built.
  8. Select the best model and activities you would like to take into the production phase.