regression: Customized all-in-one regression function

In sfeuerriegel/ResearchGroupTools: Collection of Custom Helper Functions

Description

Function performs default regression via ordinary least squares. It also supports dummy variables which are not included in the dataset data, but in a global variable attached to a formula. With this input, this function can filter for a subset, remove outliers at a certain cutoff and remove dummies that are NA.

Usage

regression(formula, data = NULL, subset = NULL, dummies = NULL,
  cutoff = NULL, rmDummyNA = TRUE, vcov = NULL)

Arguments

formula	of type formula.
data	An optional data frame contain the variables in the model (excluding the dummy variables).
subset	Vector of integers or booleans defining the subset of observations to be used.
dummies	String denoting the name of the variable (i.e. matrix or data frame) containing all dummy variables.
cutoff	Relative cutoff on each side in percent (default: NULL). Values are given in percient, e.g. 0.5 represents 0.5% at each end).
rmDummyNA	Boolean indicating whether to remove dummy variables with NA coefficient (default: removal).
vcov	Estimator used for computing the covariance matrix. Default is NULL which results in to the ordinary least squares estimator. Alternatives, are, for instance, NeweyWest from the sandwich package.

Examples

x <- 1:100
clusters <- rep(c(1, 2), 50)
dummies <- model.matrix(~ clusters)
y <- x + clusters + rnorm(100)
d <- data.frame(x = x, y = y)

m <- regression(formula("y ~ x + dummies"), data = d, subset = 1:90,
                dummies = "dummies", cutoff = 0.5)
summary(m)

library(sandwich)
m <- regression(formula("y ~ x + dummies"), data = d, subset = 1:90,
                dummies = "dummies", cutoff = 0.5, vcov = NeweyWest)
summary(m)

sfeuerriegel/ResearchGroupTools documentation built on May 29, 2019, 8:01 p.m.