P81Machine LearningBeginnerChoose8 min

Choose a Baseline Before Training a Complex Ensemble

Choose a defensible simple baseline before evaluating a complex ensemble.

#choose#baseline#model selection

Scenario

A team wants to start a customer-churn project with a large stacking ensemble, but has not measured a majority-class, logistic-regression, or simple-tree baseline.

Problem statement

Choose an appropriate first baseline and explain what it tells the team.

Your Task

  1. Select a baseline suited to the task.
  2. Explain the reference it creates.
  3. State why the ensemble may still be worth testing later.

Related Concepts

BaselinesModel selectionEvaluation

Ready to count this problem?

Mark it complete when you have reasoned through the solution in your own words.