OTrecod (version 0.1.2)

transfo_target: transfo_target()

Description

This function prepares the encoding of the target variable before running an algorithm using optimal transportation theory.

Usage

transfo_target(z, levels_order = NULL)

Value

The list returned is:

NEW

an object of class factor of the same length as z

LEVELS_NEW

the levels (ordered or not) retained for z

Arguments

z

a factor variable (ordered or not). A variable of another type will be, by default, convert to a factor.

levels_order

a vector corresponding to the values of the levels of z. When the target is ordinal, the levels can be sorted by ascending order. By default, the initial order is remained.

Author

Gregory Guernec

Details

The function transfo_target is an intermediate function direcly implemented in the functions OT_outcome and OT_joint, two functions dedicated to data fusion (see (1) and (2) for details). Nevertheless, this function can also be used separately to assist user in the conversion of a target variable (outcome) according to the following rules:

• A character variable is converted in factor if the argument levels_order is set to NULL. In this case, the levels of the factor are assigned by order of appearance in the database.

• A character variable is converted in ordered factor if the argument levels_order differs from NULL. In this case, the levels of the factor correspond to those assigned in the argument.

• A factor stays unchanged if the argument levels_order is set to NULL. Otherwise the factor is converted in ordered factor and the levels are ordered according to the argument levels_order.

• A numeric variable, discrete or continuous is converted in factor if the argument levels_order is set to NULL, and the related levels are the values assigned in ascending order.

• A numeric variable, discrete or continuous is converted in ordered factor if the argument levels_order differed from NULL, and the related levels correspond to those assigned in the argument.

References

1. Gares V, Dimeglio C, Guernec G, Fantin F, Lepage B, Korosok MR, savy N (2019). On the use of optimal transportation theory to recode variables and application to database merging. The International Journal of Biostatistics. Volume 16, Issue 1, 20180106, eISSN 1557-4679. doi:10.1515/ijb-2018-0106

2. Gares V, Omer J (2020) Regularized optimal transport of covariates and outcomes in data recoding. Journal of the American Statistical Association. tools:::Rd_expr_doi("10.1080/01621459.2020.1775615")

See Also

compare_lists

Examples

y <- rnorm(100, 30, 10)
ynew1 <- transfo_target(y)

newlev <- unique(as.integer(y))
ynew2 <- transfo_target(y, levels_order = newlev)
newlev2 <- newlev[-1]
ynew3 <- transfo_target(y, levels_order = newlev2)

outco <- c(rep("A", 25), rep("B", 50), rep("C", 25))
outco_new1 <- transfo_target(outco, levels_order = c("B", "C", "A"))
outco_new2 <- transfo_target(outco, levels_order = c("E", "C", "A", "F"))
outco_new3 <- transfo_target(outco)

outco2 <- c(rep("A", 25), NA, rep("B", 50), rep("C", 25), NA, NA)
gg <- transfo_target(outco2)
hh <- transfo_target(outco2, levels_order = c("B", "C", "A"))