gd_outsample: gradient descent using out of sample penalty

In carolineying/bis557: BIS 557 Assignment

gradient descent using out of sample penalty

Description

This is a function taking a formula, a dataframe and an optional list of contrast for factor variables, and using gradient descent with out of sample penalty and returning OLS coefficients. Large dataset should have more accurate results for splitting samples.

Usage

gd_outsample(formula, df, contrasts = NULL, gamma = 1e-04, iter = 10^6)

Arguments

formula	an object of class "formula": describing the model to be fitted.
df	a dataframe that containing all the variables in the model.
contrasts	an optional list of contrasts for factor variables.
gamma	gamma_k, the learning rate
iter	number of iterations

Examples

data(iris)
form <- Sepal.Length ~ Sepal.Width
gd_outsample(formula = form,df = iris)

carolineying/bis557 documentation built on April 17, 2023, 4 a.m.