Calculate Correlation Coefficients Between Inputs and Outputs

Computes correlation coefficients between mcmodule inputs and outputs using tornado analysis (from the mc2d package). Supports multiple correlation methods and captures warnings generated during calculation.

Usage

mcmodule_corr(
  mcmodule,
  output = NULL,
  by_exp = FALSE,
  match_variates = TRUE,
  variates_as_nsv = FALSE,
  print_summary = TRUE,
  progress = FALSE,
  method = c("spearman", "kendall", "pearson"),
  use = "all.obs",
  lim = c(0.025, 0.975)
)

Arguments

mcmodule: (mcmodule object). Module containing simulation results.
output: (character, optional). Output node name. If NULL (default), uses the last node in mcmodule$node_list. If by_exp = TRUE, uses the last output node per expression. Default: NULL.
by_exp: (logical). If TRUE, calculate correlations by expression output; if FALSE, use global output (last node). Default: FALSE.
match_variates: (logical). If TRUE, match input nodes to output variates when data dimensions differ. Default: TRUE.
variates_as_nsv: (logical). If TRUE, combine all variates into one mc object; if FALSE, analyse each variate separately. See mcmodule_to_mc(). Default: FALSE.
print_summary: (logical). If TRUE, print correlation analysis summary. Default: TRUE.
progress: (logical). If TRUE, print progress information while running. Default: FALSE.
method: (character). Correlation coefficient type: "spearman" (default), "kendall", or "pearson". See stats::cor(). Default: "spearman".
use: (character). Method for handling missing values: "all.obs", "complete.obs", or "pairwise.complete.obs". See stats::cor(). Default: "all.obs".
lim: (numeric vector). Quantiles for credible interval computation (reserved for two-dimensional models). Default: c(0.025, 0.975).

Value

A data frame with correlation coefficients and metadata. Columns include:

exp: Expression name
exp_n: Expression number
variate: Variate number
output: Output node name
input: Input node name
value: Correlation coefficient value
strength: Qualitative strength of association (Very strong, Strong, Moderate, Weak, None)
method: Correlation method used (spearman, kendall, or pearson)
use: Method for handling missing values (passed to the correlation function)
warnings: Any warnings generated during correlation calculation (if present)
Additional columns for global keys (e.g., pathogen, origin)

Examples

mcmodule <- agg_totals(
  mcmodule = imports_mcmodule,
  mc_name = "no_detect_a",
  agg_keys = "pathogen"
)
#> 3 variates per group for no_detect_a
cor_results <- mcmodule_corr(mcmodule)
#> 
#> === Correlation Analysis Summary ===
#> 
#> Analysis Parameters:
#> - Analysis type: Global output
#> - Output node: no_detect_a_agg
#> - Correlation method(s): spearman
#> - Missing value handling: all.obs
#> 
#> Expression Information:
#> - Module: mcmodule
#>   - Expression: imports
#>     - Input nodes: w_prev, test_sensi
#>     - Variates analyzed: 6
#> 
#> Results Summary:
#> - Total correlations calculated: 12
#> - Top 2 most influential inputs (by absolute mean correlation):
#>   1. w_prev: 0.4780
#>   2. test_sensi: -0.0335
#> 
#> Input Correlation Strength Distribution:
#> - Very strong: 2 (16.7%)
#> - Moderate: 1 (8.3%)
#> - Weak: 1 (8.3%)
#> - None: 8 (66.7%)
#> 
#> Inputs by Correlation Strength:
#> - Very strong: w_prev
#> - Moderate: w_prev
#> - Weak: w_prev
#> - None: test_sensi, w_prev

# Use single method
cor_results_spearman <- mcmodule_corr(mcmodule, method = "spearman")
#> 
#> === Correlation Analysis Summary ===
#> 
#> Analysis Parameters:
#> - Analysis type: Global output
#> - Output node: no_detect_a_agg
#> - Correlation method(s): spearman
#> - Missing value handling: all.obs
#> 
#> Expression Information:
#> - Module: mcmodule
#>   - Expression: imports
#>     - Input nodes: w_prev, test_sensi
#>     - Variates analyzed: 6
#> 
#> Results Summary:
#> - Total correlations calculated: 12
#> - Top 2 most influential inputs (by absolute mean correlation):
#>   1. w_prev: 0.4780
#>   2. test_sensi: -0.0335
#> 
#> Input Correlation Strength Distribution:
#> - Very strong: 2 (16.7%)
#> - Moderate: 1 (8.3%)
#> - Weak: 1 (8.3%)
#> - None: 8 (66.7%)
#> 
#> Inputs by Correlation Strength:
#> - Very strong: w_prev
#> - Moderate: w_prev
#> - Weak: w_prev
#> - None: test_sensi, w_prev