doex: An R Package for The One-Way Heteroscedastic ANOVA Tests

Installation

To install the package from the CRAN, the package can be installed by using following code:

install.packages("doex")
library(doex)

Usage

doex provides 25 statistical test functions, two datasets and an outlier generation function:

• AF(), AG(), AGF(), B2(), BF(), BX(), CF(), FA(), GF(), JF(), MBF(), MW(), OS(), OSR(), PB(), PF(), SS(), WA(), and WE() are used for heteroscedastic one-way ANOVA setup.
• MGF() and RGF() are used for heteroscedastic one-way ANOVA under non-normality caused by outliers and skewness, respectively.
• fa_exp(), gpv_exp(), and pb_exp() are used for testing equality of two-parameter distributed means under unequal scale parameters.
• component and hybrid are the datasets can be used for the usage examples of the tests.
• outly() is an outlier generation algorithm.

The Alexander-Govern (AG) test is performed with the inputs are the observations and the group labels of the observations as follows:

library(doex)
AG(hybrid$data, hybrid$species)

The modified generalized F-test is performed with the inputs are the observations and the group labels of the observations as follows:

library(doex)
MBF(hybrid$data, hybrid$species)

The parametric bootstrap test for two-parameter exponential distributed populations is performed with the inputs are the observations, the group labels of the observations and the repetition as follows:

library(doex) 
fa_exp(component$lifetime, component$supplier)

To generate outliers based on interquartile range approach, outly() function is used with some options as follows:

library(doex) 
outly(8, 2, 2, 0.05, FALSE)

Citation

If you use doex, please cite it:

@article{cavus2021,
  author = {Mustafa Cavus and Berna Yazıcı},
  title = {{Testing the equality of normal distributed and independent
            groups' means under unequal variances by doex package}},
  year = {2021},
  journal = {{The R Journal}},
  doi = {10.32614/RJ-2021-008},
  url = {https://doi.org/10.32614/RJ-2021-008},
  pages = {134--154},
  volume = {12},
  number = {2}
}

Contact

For any questions and feedback, please dont hesitate to contact me via following e-mail address:

📧 [email protected]