StrctTerm.scale_penalty()

StrctTerm.scale_penalty()

Scale the penalty matrix by its infinity norm.

The penalty matrix is divided by its infinity norm (max absolute row sum) so that its values are numerically well-conditioned for downstream use. The updated penalty replaces the previous one.

Return type:

Self

Returns:

The modified term instance (self).

See also

Basis.scale_penalty

The term calls this method internally. More details are documented there.

Examples

>>> x = jnp.linspace(0.0, 1.0, 4)
>>> basis = Basis(
...     jnp.column_stack([jnp.ones_like(x), x]),
...     xname="x",
...     penalty=2.0 * jnp.eye(2),
... )
>>> term = StrctTerm.f(basis, scale=1.0).scale_penalty()
>>> float(jnp.max(term.basis.penalty.value))
1.0