This class represents the elementwise Huber function, \(Huber(x, M) = \)
\(2M|x|-M^2\)for \(|x| \geq |M|\)
\(|x|^2\)for \(|x| \leq |M|.\)
Huber(x, M = 1)# S4 method for Huber
validate_args(object)
# S4 method for Huber
to_numeric(object, values)
# S4 method for Huber
sign_from_args(object)
# S4 method for Huber
is_atom_convex(object)
# S4 method for Huber
is_atom_concave(object)
# S4 method for Huber
is_incr(object, idx)
# S4 method for Huber
is_decr(object, idx)
# S4 method for Huber
get_data(object)
# S4 method for Huber
graph_implementation(object, arg_objs, size,
data = NA_real_)
A positive scalar value representing the threshold. Defaults to 1.
A list of arguments to the atom.
An index into the atom.
A list of linear expressions for each argument.
A vector with two elements representing the size of the resulting expression.
A list of additional data required by the atom.
validate_args: Check that M is a non-negative constant.
to_numeric: The Huber function evaluted elementwise on the input value.
sign_from_args: The atom is positive.
is_atom_convex: The atom is convex.
is_atom_concave: The atom is not concave.
is_incr: A logical value indicating whether the atom is weakly increasing.
is_decr: A logical value indicating whether the atom is weakly decreasing.
get_data: A list containing the parameter M.
graph_implementation: The graph implementation of the atom.