pmx_endpoint: Creates pmx endpoint object

In ggPMXdevelopment/ggPMX: 'ggplot2' Based Tool to Facilitate Diagnostic Plots for NLME Models

Creates pmx endpoint object

Description

Creates pmx endpoint object

Usage

pmx_endpoint(code, label = "", unit = "", file.code = code, trans = NULL)

Arguments

code	character endpoint code : used to filter observations DVID==code.
label	character endpoint label: used to set title and axis labels
unit	character endpoint unit : used to set title and axis labels
file.code	character endpoint file code : used to set predictions and finegrid files extensions in case using code parameter is not enough.
trans	list Transformation parameter not used yet.

Details

In case of multiple endpoints, pkpd case for example, we need to pass endpoint to the pmx call. Internally , ggPMX will filter the observations data set to keep only rows satisfying DVID==code. The code is also used to find the right predictions and or fingrid files. ggPMX use the configuration file to fine the path of the predictions file (like the single endpoint case) and then filter the right file using the code parameter.
For example:

• predictions{code}.txt for mlx16

• predictions{code}.txt and y{code}_residual for mlx18

For some tricky examples the code parameter is not enough to find the files. In that case the file.code parameter is used to distinguish the endpoint files.

Examples

## Use file.code parameter
pk_pd_path <- file.path(system.file(package = "ggPMX"), "testdata","pk_pd")

WORK_DIR <- file.path(pk_pd_path, "RESULTS")

ep <- pmx_endpoint(
  code="4",
  file.code="2"
)

input_file <- file.path(pk_pd_path, "pk_pd.csv")

ctr <- pmx_mlx(
  config = "standing",
  directory = WORK_DIR,
  input = input_file,
  dv = "dv",
  dvid = "dvid",
  cats = "sex",
  conts = "wt",
  endpoint = ep
 )

## using mlxtran 

ep <- pmx_endpoint(
  code="3",
 file.code="1"
)

mlxtran_file <- file.path(pk_pd_path, "pk_pd.mlxtran")
ctr <- pmx_mlxtran(mlxtran_file,endpoint=ep)

ggPMXdevelopment/ggPMX documentation built on Dec. 11, 2023, 5:24 a.m.