Module: ci_rvm.ci_rvm_mpinv


find_CI_bound(index, target, x0, fun, jac, hess)

find_profile_CI_bound(index, direction, x0, …)

find_profile_CI_bound_steps(index, x0, fun, …)

is_negative_semidefinite(M[, tol, …])

find_CI_bound(index: index of the parameter to consider, target: target log-likelihood l*, x0: maximum likelihood estimator (MLE), fun: log-likelihood function, jac: gradient of fun, hess: hessian of fun, forward: True, if right end point of CI is sought, else False = True, fun0: log-likelihood at MLE = None, jac0: gradient of log-liekelihood at MLE = None, hess0: Hessian of log-likelihood at MLE = None, nmax: maximal number of iterations = 200, nchecks: maximal number of trust-region changes per iteration = 65, apprxtol: relative tolerance between f and its approximation = 0.5, resulttol: tolerance of the result (f and norm(jac)) = 0.001, singtol: tolerance for singularity checks = 0.0001, minstep: controls the minimal radius of the trust region = 1e-05, radiusFactor: In [1, 2]. Controls how quickly the trust region decreases = 1.5, infstep: Stepsize after which a parameter is deemed unestimbale = 10000000000.0, maxRadius: radius of the trust region in the last iteration = 10000.0, disp: whether to print a status message in each iteration = True, track_x: whether to return the parameter trace = False, track_f: whether to return the log-likelihood trace = False)[source]
find_profile_CI_bound(index, direction, x0, fun, jac, hess, alpha=0.95, fun0=None, *args, **kwargs)[source]
find_profile_CI_bound_steps(index, x0, fun, jac, hess, direction=1, alpha=0.95, stepn=1, fun0=None, hess0=None, nmax=200, epsilon=0.0001, disp=True, vm=False)[source]
is_negative_semidefinite(M, tol=1e-06, return_singular=False)[source]