README.md

In ppanko/orthogonalize: Estimate OLS residuals, simply and quickly

orthogonalize

The aim of the orthogonalize package is to streamline the process of residualizing covariates by letting C++ do some of the heavy lifting.

Install orthogonalize in your R session by running the following code:

devtools::install_github("ppanko/orthogonalize")

To check the help documentation in R:

?orthogonalize

## OR

help(orthogonalize)

 

Usage

For ease of access, orthogonalize provides a formula interface and a few sane defaults. Basic functionality can look like this:

library(orthogonalize)

data(iris)

petalWidthPrime <- orthogonalize(
  Petal.Width ~ ., iris 
)

 

Slightly more complex requirements may call for within-group residuals or residuals offset by the intercept:

## By group 
pwPrimeGroup <- orthogonalize(
  formula = Petal.Width ~ ., 
  data    = iris,
  group   = "Species"
)

## Add intercept
pwPrimeIntercept <- orthogonalize(
  formula   = Petal.Width ~ ., 
  data      = iris,
  intercept = TRUE
)

 

Following the main idea of letting C++ do the work, users can also supply multiple formulas in a list, as opposed to writing loops in R:

## Multiple formulas
widthPrimeMat <- orthogonalize(
  formula   = list(
                Petal.Width ~ ., 
        Sepal.Width ~ . 
          ),
  data      = iris
)

 

For a more modular apporach, the intercept and group arguments in the multi-formula case can either take on single values (the default) as in the example above or be tweaked to suit user requirements:

## Add intercept to first residual vector, 
## but not the second.
widthPrimeInt <- orthogonalize(
  formula   = list(
                Petal.Width ~ ., 
        Sepal.Width ~ . 
          ),
  data      = iris,
  intercept = c(TRUE, FALSE)
)

## Single-group residuals for first residual
## vector and within-group residuals for the
## second. 
widthPrimeGrp <- orthogonalize(
  formula   = list(
                Petal.Width ~ ., 
        Sepal.Width ~ . 
          ),
  data      = iris,
  group     = c("", "Species")
)

## Tweaking both intercept and group 
## parameters for each residual 
## vector.
widthPrimeMat <- orthogonalize(
  formula   = list(
                Petal.Width ~ ., 
        Sepal.Width ~ . 
          ),
  data      = iris,
  intercept = c(TRUE, FALSE),
  group     = c("Species", "")
)

If you encounter any bugs while using this software or have a suggestion, please file an issue or pull request on this page. For more general questions or assistance, please email me at [e-mail].

ppanko/orthogonalize documentation built on July 17, 2020, 2:24 a.m.