nn_sgd: Stochastic gradient descent (SGD) to estimate a neural net

In casxue/bis557: Coursework for BIS 557

Description

Apply stochastic gradient descent (SGD) to estimate NN. Modified from CASL.

Usage

nn_sgd(X, y, sizes, epochs, eta, weights = NULL, f_p = util_mse_p)

Arguments

X	a numeric matrix
y	a numeric vector
sizes	an integer vector
epochs	an integer value
eta	a positive numeric value
weights	optional list of starting weights
f_p	derivative of the loss function

Value

a list of trained weights

casxue/bis557 documentation built on May 7, 2019, 5 a.m.