detect_outliers: Outliers detection

In pnolain/pmxploit: Pharmacometric tools for post-processing of mixed-effects models for continuous data

Outliers detection

Description

Detection of observations outliers based on residuals values according to two different methods: Grubb's test or a non-parametric method.

Usage

detect_outliers(
  run,
  compartment = NULL,
  residuals = NULL,
  method = "grubbs",
  grubbs_pvalue_threshold = 0.05,
  boxplot_coefficient = 3,
  keep_time_zero = FALSE
)

Arguments

run	pmxploit NONMEM run object.
compartment	integer. Number of the compartment of the observations of interest.
residuals	character. Column name of the residuals in the output tables.
method	character. One of "grubbs" (for iterative Grubb's test) or "boxplot" (for non-parametric detection).
grubbs_pvalue_threshold	numeric. p-value threshold for Grubb's test.
boxplot_coefficient	numeric. k coefficient for non-parametric test.
keep_time_zero	logical. If TRUE, does not ignore values at initiation (when TIME = 0). Default is FALSE.

Value

A list with the following structure:

• method: character string of outlier detection method.

• residuals: character string of the type of residuals.

• source: tibble of the data source.

• outliers: tibble of the outliers with 5 columns: ID, TIME, CMT, DV and "residuals".

Examples

EXAMPLERUN %>%
 detect_outliers(compartment = 2, residuals = "CWRES", method = "grubbs")

EXAMPLERUN %>%
 detect_outliers(compartment = 2, residuals = "CWRES",
                 method = "grubbs", grubbs_pvalue_threshold = 0.10)

EXAMPLERUN %>%
 detect_outliers(compartment = 2, residuals = "CWRES", method = "boxplot")

pnolain/pmxploit documentation built on Jan. 31, 2024, 1:16 p.m.