Close the boundary: this is an explicit preprocessing rewrite, not a model claim.
highlighted = computed this step
What missing-value handling does
It rewrites displayed missing cells using the pinned preprocessing policy so later methods receive explicit values and indicators.
missing cells→filled cells plus indicators
What it does not do
It does not train a model and does not prove accuracy, calibration, model quality, or probability truth.
preprocessing only
The honest boundary
This book pins one missing-value policy. It is NOT training, NOT learning, NOT generalization, NOT accuracy, NOT calibration, NOT a guarantee of model quality, and NOT probability truth.