mixle.utils.automatic.detectors.skew_normal module¶
Skew-normal continuous candidate – a Gaussian tilted by an asymmetry (shape) parameter.
The skew-normal only earns its third parameter when the data carry a genuine, in-family asymmetry: the
gate requires a non-trivial sample skewness whose magnitude stays inside the family’s reach (the
skew-normal can only attain |skewness| < ~0.995), so symmetric Gaussian data and heavily-skewed
exponential data both fall outside the gate and are left to the builtins.