Many Welch tests.
Many Welch tests.
welch.tests(y, x, logged = FALSE, parallel = FALSE)
y |
The dependent variable, a numerical vector. |
x |
A matrix with the indendent variables. They must be integer valued data starting from 1, not 0 and be consecutive numbers. Instead of a data.frame with factor variables, the user must use a matrix with integers. |
logged |
Should the p-values be returned (FALSE) or their logarithm (TRUE)? |
parallel |
If you want to run the function in parallel set this equal to TRUE. |
For each categorical predictor variable, a Welch test is performed. This is useful in feature selection algorithms, to determine for which variable, the means of the dependent variable differ across the different values.
A two-column matrix with the test statistics (F test statistic) and their associated p-values (or their loggarithm).
Michail Tsagris
R implementation and documentation: Michail Tsagris <mtsagris@uoc.gr>
B.L. Welch (1951). On the comparison of several mean values: an alternative approach. Biometrika, 38(3/4), 330-336.
y <- rnorm(200) x <- matrix(rbinom(200 * 50, 2, 0.5), ncol = 50) + 1 a <- welch.tests(y, x)
Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.