get_trt: get_trt

In TSDT: Treatment-Specific Subgroup Detection Tool

get_trt

Description

Returns the treatment variable in the in-bag or out-of-bag data.

Usage

get_trt(data, scoring_function_parameters = NULL)

Arguments

data	A data.frame containing in-bag or out-of-bag data
scoring_function_parameters	A list of named elements containing control parameters and other data required by the scoring function

Details

If the user provides a trt_var parameter in the list of scoring_function_parameters this function will return the variable specified by that parameter. If the user specifies a trt_col parameter in the list of scoring_function_parameters the function returns the column in data indexed by that parameter. Lastly, if data contains a variable called 'trt' that variable is returned. Otherwise, NULL is returned.

Value

Treatment variable (if available) or NULL.

See Also

get_y, get_covariates

Examples

## Create an example data.frame
df <- data.frame( y <- 1:5 )
names( df ) <- "y"
df$time <- 10:14
df$time2 <- 20:24
df$event <- sample( c(0:1), size = 5, replace = TRUE )
df$trt <- sample( c("Control","Treatment"), size = 5, replace = TRUE )
df$x1 <- runif( n = 5 )
df$x2 <- LETTERS[1:5]

## Select the trt variable by name
get_trt( df, scoring_function_parameters = list( trt_var = 'trt' ) )

## Select the trt variable by column index
get_trt( df, scoring_function_parameters = list( trt_col = 5 ) )

## The default behavior works for this example because the trt variable in df
## is actually called trt.
get_trt( df )

## If the user's data does not contain a variable called
## 'y' the default behavior will fail. In this case the user must explicitly
## identify the 'y' variable via one of the two previous methods.
names( df )[which(names(df) == "trt")] <- "treatment" # rename the 'trt' variable to 'treatment'

get_trt( df )  # now default behavior fails (i.e. returns NULL)

get_trt( df, scoring_function_parameters = list( trt_var = 'treatment' ) ) # this works

TSDT documentation built on April 7, 2022, 1:07 a.m.