pub fn zscore(values: &[f64]) -> Vec<f64>Expand description
Z-score normalisation against the partition’s own mean and sample standard deviation.
The statistics are taken over the non-NULL values; a NULL row stays NULL and does not make the rest of the partition NULL, which is what summing a slice containing one NaN used to do. Returns all-NAN when fewer than two values are present or the spread is zero.
Uses the sample standard deviation (Bessel’s correction, n-1) for
consistency with crate::compute::simd_std_dev.