configurations: Default Configurations for Model Estimation

In itamarfaran/corrpops: Analysis of populations of correlation matrices

Description

A list of default configurations for estimate_model. running the function will print the defaults.

• model_reg: a list with params lambda, lp for regularization of alpha. the loss function has an element of reg_lambda * sum(abs(alpha - LinkFunc$null_value)^reg_p) ^ (1/reg_p)

• matrix_reg: parameters passed on to regularize_matrix. the weighting matrix (covariance matrix of correlations) is regularized accordingly

• iterations: parameters used for the iterations' stopping rule. the definitions of 'maxit', 'reltol', 'abstol' are the same is in optim. iterations will stop only if there were 'minit' optim iterations with convergence=0 in a row.

• optim: possible parameters to be passed to optim, namely 'method', 'reltol', 'abstol'. also, log_optim is a boolean - whether to save the result of the call to optim in each iteration.

Usage

configurations(index)

Arguments

index

which configurations to print. if missing, print all. can be one or more of the aforementioned.

itamarfaran/corrpops documentation built on Dec. 20, 2021, 8:02 p.m.