Why might Anderson–Darling detect tail mismatch more readily than a maximum-CDF-distance test?
StatisticsLevel: IntermediateDistribution checking
Why might Anderson–Darling detect tail mismatch more readily than a maximum-CDF-distance test?
Hints
- Compare weighting across quantiles.
- A maximum and a weighted integral differ.
Answer
Its weighted discrepancy gives tails more emphasis. This makes the choice of goodness-of-fit test relevant to the kinds of departures that matter.
Source and licence: SciPy — distribution diagnostics and robust tests · BSD-3-Clause · answer: Quantoros editorial