est_kcde_params_stepwise_crossval_one_potential_step: Estimate the parameters theta and corresponding...

In reichlab/kcde: Kernel conditional density estimation with flexible kernel specifications

Description

Estimate the parameters theta and corresponding cross-validated estimate of loss for one possible model obtained by adding or removing a variable/lag combination from the model obtained in the previous iteration of the stepwise search procedure.

Usage

est_kcde_params_stepwise_crossval_one_potential_step(prev_vars_and_offsets,
  prev_theta, prev_phi, init_theta_vector, init_phi_vector, update_var_name,
  update_lag_value, data, all_na_drop_rows, kcde_control)

Arguments

prev_vars_and_offsets	list representing combinations of variables and lags included in the model obtained at the previous step
prev_theta	list representing the kernel parameter estimates obtained at the previous step
update_var_name	the name of the variable to try adding or removing from the model
data	the data frame with observations used in estimating model parameters
kcde_control	a list of parameters specifying how the fitting is done
update_offset_value	the value of the offset for the variable specified by update_var_name to try adding or removing from the model

Value

a list with three components: loss is a cross-validation estimate of the loss associated with the estimated parameter values for the given model, lags is a list representing combinations of variables and lags included in the updated model, and theta is a list representing the kernel parameter estimates in the updated model

reichlab/kcde documentation built on May 27, 2019, 4:53 a.m.