compare_cvma: Compare measures of association between two fits

In benkeser/cvma: Cross-validation-based maximal association

Description

Compare measures of association between two fits

Usage

compare_cvma(object1, object2, contrast = "diff",
  compare_control = list(f = function(h_cv_measures) {     h_cv_measures
  }, f_inv = function(h_cv_measures) {     h_cv_measures }, h =
  function(cv_measure1, cv_measure2) {     cv_measure1 - cv_measure2 },
  fh_grad = function(cv_measure1, cv_measure2) {     c(1, -1) }, null = 0),
  alpha = 0.05)

Arguments

object1	An object of class cvma
object2	An object of class cvma
contrast	A character indicating what contrast to take. Can be "diff", "ratio", or "logratio".
compare_control	Contains information regarding format of the contrast we are interested in, its inverse function, type of contrast, IC components of the corresponding covariance calculation, and null hypothesis. Defaults to functions necessary for "diff".
alpha	Confidence intervals of 1 - alpha.

Value

Estimates, CI, and p-value.

Examples

set.seed(1234)
library(SuperLearner)
library(future)
X <- data.frame(x1=runif(n=100,0,5), x2=runif(n=100,0,5))
Y1 <- rnorm(100, X$x1 + X$x2, 1)
Y2 <- rnorm(100, X$x1 + X$x2, 3)
Y <- data.frame(Y1 = Y1, Y2 = Y2)
fit1 <- cvma(Y = Y, X = X, V = 5, 
                learners = c("SL.glm","SL.mean"))
fit2 <- cvma(Y = Y, X = X[,-2,drop = FALSE], V = 5, 
                learners = c("SL.glm","SL.mean"))
# difference in R^2
compare_cvma(fit1, fit2, "diff")
# ratio of R^2, symmetric CI
compare_cvma(fit1, fit2, "ratio")
# ratio of R^2, CI symmetric on log-scale
compare_cvma(fit1, fit2, "logratio")

benkeser/cvma documentation built on May 5, 2019, 1:37 p.m.