sim.my: Simulate data

In chngpt: Estimation and Hypothesis Testing for Threshold Regression

Simulate data

Description

Simulate data

Usage

sim.my(n, seed, label, alpha, beta, e. = NULL, b. = NULL, tr. = NULL)

Arguments

n	Sample size
seed	Seed for random number generator
label	A character string which specifies the simulation scenario. sigmoid4, sigmoidgam4, elbow4
alpha	regression parameter
beta	regression parameter
e.	inflection point for the logistic transformation (the log scale)
b.	slope for the logistic transformation
tr.	threshold point

Details

When the label starts with elbow, the transformation on x.star is elbow shaped. When the label starts with sigmoid, the transformation on x.star is sigmoid shaped. Data simulated from logit(Pr(Y==1))=alpha + beta*(transformed x.star).

Value

A data frame with columns: y, x.star, x.star.expit (if label starts with sigmoid), x.star.tr (if label starts with elbow), x.bin.med (x.star dichotomized at median), x.tri (x.star trichotomized at tertiles).

Examples

alpha=-1; beta=log(0.2)
e.=5; b.=-30; t.=1
dat=sim.my(n=250, seed=1, label="sigmoid4", alpha, beta, e.=e., b.=b.)

chngpt documentation built on Feb. 16, 2023, 10:38 p.m.