el0ps.penalty.compute_param_slope_neg

el0ps.penalty.compute_param_slope_neg(penalty, i, lmbd, tol=1e-08, maxit=100)

Utility to approximate the value of the function penalty.param_slope_neg for a given BasePenalty instance using a bisection method.

Return type:

float

Parameters:

penaltyBasePenalty

The penalty function.

iint

Parameter involved in the penalty.param_slope_neg function.

lmbdfloat

Parameter involved in the penalty.param_slope_neg function.

tolfloat = 1e-4

Bisection approximation tolerance.

maxitint = 100

Maximum number of bisection iterations.