optimize (pmrf.optimize)

Non-linear optimization of RF models.

Provides solvers and routines to find the optimal point-estimates that minimize a given objective/cost function.

Classes

AbstractBoundedMinimizer()	Abstract interface for bounded minimization algorithms.
AbstractUnconstrainedMinimizer()	Abstract interface for unconstrained minimization algorithms.
OptimizeResult(model, objective, success[, ...])	The result of an optimization run.
ScipyMinimize(method, tol, options, ...)	A wrapper around SciPy's scipy.optimize.minimize().
OptimistixMinimise(solver[, fatol])	An optimizer that wraps an arbitrary Optimistix optimistix.AbstractMinimiser.
LBFGSB([gtol, stepsize, linesearch])	A L-BFGS-B optimizer in JAX.
NelderMead(fatol, xatol, xrtol, norm, ...)	A Nelder-Mead optimizer in JAX.
LBFGS(fatol, step_atol, step_rtol, norm, ...)	A LBFGS optimizer in JAX.
BFGS(fatol, step_atol, step_rtol, norm, ...)	A BFGS optimizer in JAX.
GradientDescent(learning_rate, fatol, ...)	A Gradient Descent optimizer in JAX.

Functions

is_optimizer(x)	Returns True if pmrf.optimize.is_minimizer returns True.
is_minimizer(x)	Returns True if x is an instance of pmrf.optimize.AbstractUnconstrainedMinimizer or pmrf.optimize.AbstractBoundedMinimizer.
minimize(objective, model, frequency[, ...])	Minimizes a given objective function for a model over a frequency range.