osm_gradient_descent: out-of-sample gradient descent

In Z1chenZhao/bis557: BIS557

Description

Implement gradient descent for ordinary least squares

Usage

osm_gradient_descent(
  form,
  dat,
  contrasts = NULL,
  factor = 1e-04,
  m.iteration = 1e+05,
  b.difference = 1e-10,
  F = 10
)

Arguments

form	A formula with the format of "Y ~ .".
dat	A dataframe.
contrasts	A list of contrasts.
factor	The factor to adjust beta, set default as 0.0001.
m.iteration	The maximum iteration times to allow the loop to exit, set default as 1e5
b.difference	The difference between previous sum of residuals and new sum of residuals to allow the loop to exit, set default as 1e-10.
F	The number of folds, set default as 10

Value

A list of coefficients from an linear regression model with gradient descent.

Examples

## Not run: 
osm_gradient_descent(Sepal.Length ~ ., iris, contrasts)

## End(Not run)

Z1chenZhao/bis557 documentation built on Dec. 21, 2020, 10:03 p.m.