Generate Simulated Data with Multinomial Logistic Regression Model — gendata_MLR • HTLR

This function generates the response variables y given optional supplied X using a multinomial logistic regression model.

Usage

gendata_MLR(n, p, NC = 3, nu = 2, w = 1, X = NULL, betas = NULL)

Arguments

n: Number of observations.
p: Number of features.
NC: Number of classes for response variables.
nu, w: If betas is not supplied (default), the regression coefficients are generated with t prior with df = nu, scale = sqrt(w); will be ignored if betas is supplied.
X: The design matrix; will be generated from standard normal distribution if not supplied.
betas: User supplied regression coefficients.

Value

A list contains input matrix X, response variables y, and regression coefficients deltas.

See also

Examples

set.seed(12345)
dat <- gendata_MLR(n = 100, p = 10)
ggplot2::qplot(dat$y, bins = 6)

corrplot::corrplot(cor(dat$X))