glm_adapt: Gradient Descent for GLM (adaptive step)

In yijunyang/bis557: BIS557 HW#1

Description

Implement a first-order solution for the GLM maximum likelihood problem using only gradient information, avoiding the Hessian matrix.

Usage

glm_adapt(
  X,
  Y,
  mu_fun,
  var_fun,
  maxit = 1e+06,
  tol = 1e-10,
  gamma = 1e-05,
  m = 0.9
)

Arguments

X	design matrix
Y	response vector
mu_fun	function from eta to the expected value
maxit	integer maximum number of iterations
tol	numeric tolerance parameter
gamma	The step size
mom	momentum parameter

Value

a list of beta coefficients

Author(s)

Yijun Yang

Examples

## Not run: 
glm_adapt(X, Y, mu_fun = function(eta) 1/(1+exp(-eta)), var_fun = function(eta) eta)

## End(Not run)

yijunyang/bis557 documentation built on Dec. 21, 2020, 3:06 a.m.