RCIT: RCIT and RCoT - tests whether x and y are conditionally...

In ericstrobl/RCIT: The Randomized Conditional Independence Test (RCIT) and the Randomized conditional Correlation Test (RCoT)

Description

RCIT and RCoT - tests whether x and y are conditionally independent given z. Calls RIT if z is empty.

Usage

RCIT(x, y, z = NULL, approx = "lpd4", corr = TRUE, num_f = 25,
  seed = NULL)

Arguments

x	Random variable x.
y	Random variable y.
z	Random variable z.
approx	Method for approximating the null distribution. Options include: "lpd4," the Lindsay-Pilla-Basak method (default), "gamma" for the Satterthwaite-Welch method, "hbe" for the Hall-Buckley-Eagleson method, "chi2" for a normalized chi-squared statistic, "perm" for permutation testing (warning: this one is slow but recommended for small samples generally <500 )
corr	TRUE corresponds to RCoT and FALSE to RCIT. Default is TRUE.
num_f	Number of features for conditioning set. Default is 25.
seed	The seed for controlling random number generation. Use if you want to replicate results exactly. Default is NULL.

Value

A list containing the p-value p and statistic Sta

Examples

RCIT(rnorm(1000),rnorm(1000),rnorm(1000));

x=rnorm(10000);
y=(x+rnorm(10000))^2;
z=rnorm(10000);
RCIT(x,y,z,seed=2);

ericstrobl/RCIT documentation built on June 5, 2019, 9:07 a.m.