Skip to contents

Evaluate effect size and significance of outliers on R

eval_diag.Rd

eval_diag function for the evaluation of effect size and significance of outliers on R detected with diagnostics, such as Cook's D or sigma rejection (Cameca default method).

Usage

eval_diag(
  .IC,
  .ion1,
  .ion2,
  ...,
  .nest = NULL,
  .X = NULL,
  .N = NULL,
  .species = NULL,
  .t = NULL,
  .flag = NULL,
  .execution = NULL,
  .output = "inference",
  .tf = "ppt",
  .label = "none",
  .meta = FALSE,
  .mc_cores = 1
)

Arguments

.IC

A tibble containing ion count data and diagnostics generated with diag_R(), as a minimum the outlier flag variable is required.

.ion1

A character string constituting the rare isotope (e.g. "13C").

.ion2

A character string constituting the common isotope (e.g. "12C").

...

Variables for grouping.

.nest

A variable hat identifies a series of analyses to calculate the significance of inter-isotope variability.

.X

A variable constituting the ion count rate (defaults to variables generated with read_IC())

.N

A variable constituting the ion counts (defaults to variables generated with read_IC().).

.species

A variable constituting the species analysed (defaults to variables generated with read_IC()).

.t

A variable constituting the time of the analyses (defaults to variables generated with read_IC()).

.flag

A variable constituting the outlier flag (defaults to variables generated with diag_R()).

.execution

A variable constituting the iterative cycles of diagnostics (defaults to variables generated with diag_R()).

.output

A character string for output as summary statistics ("inference") and statistics with the original data ("complete").

.tf

Variable transformation as parts per thousand ("ppt") or log ("log") before mixed linear model application.

.label

A character string indicating whether variable names are latex ("latex") or webtex ("webtex") compatible. Will be extended in the future default = NULL.

.meta

Logical whether to preserve the metadata as an attribute (defaults to TRUE).

.mc_cores

Number of workers for parallel execution (Does not work on Windows).

Value

A tibble::tibble() with model output. See point::names_model for more information on the model results.

Examples

# Simulated IC data
tb_dia <- diag_R(simu_IC, "13C", "12C", type.nm, spot.nm,
                 .output = "diagnostic")

# Evaluate significance and effect of outliers based on Cook's D
eval_diag(tb_dia, "13C", "12C", type.nm, spot.nm, .nest = type.nm,
          .X = Xt.pr, .N = N.pr, .species = species.nm, .t = t.nm)
#> # A tibble: 9 × 11
#>   execution type.nm    spot.nm ratio.nm M_R_Xt.pr F_R_Xt.pr p_R_Xt.pr
#>       <dbl> <chr>        <int> <chr>        <dbl>     <dbl>     <dbl>
#> 1         1 asymmetric       1 13C/12C     0.0111    63.2    1.28e-39
#> 2         1 asymmetric       2 13C/12C     0.0111    74.9    1.09e-46
#> 3         1 asymmetric       3 13C/12C     0.0111    71.9    6.70e-45
#> 4         1 ideal            1 13C/12C     0.0112     0.267  8.49e- 1
#> 5         1 ideal            2 13C/12C     0.0112     0.219  8.83e- 1
#> 6         1 ideal            3 13C/12C     0.0112     1.73   1.59e- 1
#> 7         1 symmetric        1 13C/12C     0.0110   105.     2.00e-64
#> 8         1 symmetric        2 13C/12C     0.0110   127.     2.81e-77
#> 9         1 symmetric        3 13C/12C     0.0110   122.     1.42e-74
#> # … with 4 more variables: hat_M_M_R_Xt.pr <dbl>, hat_RS_M_R_Xt.pr <dbl>,
#> #   dAIC_M_R_Xt.pr <dbl>, p_M_R_Xt.pr <dbl>