Customise Linearity Evaluation Workflow

By default the function evaluate_linearity will look for column names r_corr, pra_linear, mandel_p_val and concavity_column that represent the Correlation Coefficient, Percent Residual Accuracy and Mandel’s Fitting Test \(p\) value respectively.

It is possible to change the column names representing the three parameters to suit your own dataset using corrcoef_column, pra_column, mandel_p_val_column and concavity_column

library(lancer)

curve_name <- c("Curve_1", "Curve_2", "Curve_3", "Curve_4")
corrcoef <- c(0.952, 0.978, 0.971, 0.344)
pra <- c(65.8, 90.2, 72.9, -233)
mandel_pval <- c(0.00000029, 0.290, 0.000000032, 0.0363)
concavity <- c(-4134, -3.35, 0.394, -19.9)

curve_summary <- tibble::tibble(
  Curve_Name = curve_name,
  Corrcoef = corrcoef,
  Pra = pra,
  Mandel_Pval = mandel_pval,
  Concavity = concavity
)

curve_classified <- lancer::evaluate_linearity(
  curve_summary = curve_summary,
  grouping_variable = c(
    "Curve_Name"
  ),
  corrcoef_column = "Corrcoef",
  pra_column = "Pra",
  mandel_p_val_column = "Mandel_Pval",
  concavity_column = "Concavity"
)

print(curve_classified, width = 100)
#> # A tibble: 4 × 7
#>   Curve_Name wf1_group      wf2_group         Corrcoef    Pra Mandel_Pval
#>   <chr>      <chr>          <chr>                <dbl>  <dbl>       <dbl>
#> 1 Curve_1    Poor Linearity Saturation Regime    0.952   65.8 0.00000029 
#> 2 Curve_2    Good Linearity Good Linearity       0.978   90.2 0.29       
#> 3 Curve_3    Poor Linearity Noise Regime         0.971   72.9 0.000000032
#> 4 Curve_4    Poor Linearity Poor Linearity       0.344 -233   0.0363     
#>   Concavity
#>       <dbl>
#> 1 -4134    
#> 2    -3.35 
#> 3     0.394
#> 4   -19.9

One can change the threshold of the Correlation Coefficient (0.8), Percent Residual Accuracy (80) and Mandel’s Fitting Test \(p\) value (0.05) using corrcoef_min_threshold, pra_min_threshold and mandel_p_val_threshold respectively.

curve_classified <- lancer::evaluate_linearity(
  curve_summary = curve_summary,
  grouping_variable = c(
    "Curve_Name"
  ),
  corrcoef_column = "Corrcoef",
  corrcoef_min_threshold = 0.7,
  pra_column = "Pra",
  pra_min_threshold = 70,
  mandel_p_val_column = "Mandel_Pval",
  mandel_p_val_threshold = 0.01,
  concavity_column = "Concavity"
)

print(curve_classified, width = 100)
#> # A tibble: 4 × 7
#>   Curve_Name wf1_group      wf2_group         Corrcoef    Pra Mandel_Pval
#>   <chr>      <chr>          <chr>                <dbl>  <dbl>       <dbl>
#> 1 Curve_1    Poor Linearity Saturation Regime    0.952   65.8 0.00000029 
#> 2 Curve_2    Good Linearity Good Linearity       0.978   90.2 0.29       
#> 3 Curve_3    Good Linearity Good Linearity       0.971   72.9 0.000000032
#> 4 Curve_4    Poor Linearity Poor Linearity       0.344 -233   0.0363     
#>   Concavity
#>       <dbl>
#> 1 -4134    
#> 2    -3.35 
#> 3     0.394
#> 4   -19.9

Likewise for the excel output using the function write_summary_excel, we can choose which column to be Correlation Coefficient, Percent Residual Accuracy and Mandel’s Fitting Test \(p\) value, as well as its threshold value. The sheet name can also be changed.

lancer::write_summary_excel(
  curve_summary = curve_classified,
  file_name = "curve_summary.xlsx",
  sheet_name = "Curve Summary",
  corrcoef_column = "Corrcoef",
  corrcoef_min_threshold = 0.7,
  pra_column = "Pra",
  pra_min_threshold = 70,
  mandel_p_val_column = "Mandel_Pval",
  mandel_p_val_threshold = 0.01,
  workflow1_column = "wf1_group",
  workflow2_column = "wf2_group",
  pass_criteria_words = c("Good Linearity")
)