first_order_glm: First-Order GLM Maximum Likelihood Solution

In siqiangsu/bis557: BIS557 Homework Package

Description

First-Order GLM Maximum Likelihood Solution

Usage

first_order_glm(
  X,
  y,
  mu_fun,
  maxit = 100,
  tol = 1e-10,
  size,
  adapt = FALSE,
  gamma = 0.9
)

Arguments

X	A matrix of independent variables
y	A vector of response variable
mu_fun	The function specifying the mean function in exponential family
maxit	Maximum number of iterations
tol	Tolerance for the convergence
size	A number specifying the constant step size in gradient descent
adapt	A logical value. If TRUE, we use momentum gradient descent. If FALSE, we use constant step size
gamma	A parameter controlling the momentum update

Value

A vector of fitted regression coefficients.

Examples

## Not run: first_order_glm(X, y,mu_fun = function(eta) exp(eta), size = 0.00001,adapt = FALSE,maxit = 1000)

siqiangsu/bis557 documentation built on Dec. 21, 2020, 2:15 a.m.