Skip to content
ML Academy

Classical ML

Interview question

  • seniorlevel
  • 12 minsuggested time
  • 2Data scientist, ML engineer

A no-show model trained on a dataset where no-shows were undersampled to 50/50 is now used to compute expected no-shows per flight for overbooking. Revenue management says the expected counts are far too high. What happened, and how do you fix it?

Tests: Probability calibration, the effect of resampling on predicted probabilities, correction methods, and evaluating calibration.

day 13day 16

Sign in to practise: answers are saved before model answers are shown.