boost_gaussianLSS: Boosting GaussianLSS Model

In FAUBZhang/ASL:

Description

Fit boosting GaussianLSS model with different kinds of step lengths.

Usage

boost_gaussianLSS(
  y,
  data,
  m_stop = 1000,
  center_x = TRUE,
  weights = NULL,
  method = "SAASL"
)

Arguments

y	response variable
data	feature matrix
m_stop	stopping iteration, default is 1000
center_x	center the feature matrix? default is TRUE
weights	a vector of weights indicating the training (1) and testing (0) data, default is NULL, i.e. all observations as training data
method	step length method, should be one of "FSL", "ASL", "SAASL", "SAASL05", default is "SAASL"

Value

mu the estiamted coefficients of mean
sigma the estimated coefficients of the predictor of standard deviation
mu_mat matrix of the estimated coefficients of mu in each iteartion
si_mat matrix of the estimated coefficients of sigma in each iteration
v_mu step length of mu in each iteartion
v_mu step length of sigma in each iteration
muvsi which distribution paramter is updated in each iteration
v_mu_var which variable is used for updating mu in each iteration
v_si_var which variable is used for updating sigma in each iteartion
like_train positive likelihood of the training data in each iteartion
like_test positive likelihood of the test data in each iteration

FAUBZhang/ASL documentation built on Jan. 22, 2021, 11:46 p.m.