Closed NiallJeffrey closed 2 years ago
This discrepancy should coincide with the number of scaling moments estimates ("L" class of moments).
You can check that the combination of wph_op.scaling_moments_indices and wph_op.wph_moments_indices gives you a consistent number of coefficients with what is returned by the apply method.
When I run wph_op.wph_moments_indices(), I get a shape that doesn't match the coefficients returns by wph_op.apply().
Is this the desired behaviour?