solve_entropy_problem: Generic solver for minimum-entropy pricing kernels

In piotrek-orlowski/entRsdf: Minimum-dispersion (entropy) SDFs and their regularized approximations

Description

Given an objective function, its gradient and hessian (in the entropy_foos object), this function submits them to a convex solver and returns the vector of optimal portfolio weights plus a normalization weight in the first position. This function is supposed to be used internally in the implementation of generic methods for SDF fitting.

Usage

solve_entropy_problem(
  entropy_foos,
  excess_return_matrix,
  theta_vector_init = rep(1, ncol(excess_returns))/ncol(excess_returns),
  solver_trace = FALSE,
  ...
)

Arguments

entropy_foos	object of entropy_functions S4 class
excess_return_matrix	T x N matrix of excess returns
theta_vector_init	(N+1) x 1 numeric or \cidematrix of initial weights
solver_trace	sets the trace argument in cccp::ctrl call which is passed to cccp::cccp
...	arguments passed to

piotrek-orlowski/entRsdf documentation built on Nov. 3, 2021, 6:47 a.m.