Skip to contents

Monte Carlo Simulation Convergence Analysis

Source: R/sensitivity_analysis.R
mcmodule_converg.Rd

Analyzes convergence in Monte Carlo simulations by computing statistical measures across iterations. Calculates both standardized and raw differences between consecutive iterations to evaluate stability and convergence.

Usage

mcmodule_converg(
  mcmodule,
  from_quantile = 0.95,
  to_quantile = 1,
  conv_threshold = NULL
)

Arguments

mcmodule

A Monte Carlo module object containing simulation results

from_quantile

Lower bound quantile for analysis (default: 0.95)

to_quantile

Upper bound quantile for analysis (default: 1)

conv_threshold

Optional custom convergence threshold for standardized differences

Value

A data frame with convergence statistics per node:

  • expression: Expression identifier

  • variate: Variate identifier

  • node: Node identifier

  • max_std_dif: Max of standardized differences

  • max_dif: Max of raw differences

  • conv_threshold: Convergence at custom threshold (TRUE/FALSE), if threshold provided

  • conv_01: Convergence at 1% std threshold (TRUE/FALSE)

  • conv_025: Convergence at 2.5% std threshold (TRUE/FALSE)

  • conv_05: Convergence at 5% std threshold (TRUE/FALSE)

Details

The function performs the following:

  • Calculates convergence statistics for specified quantile range

  • Generates diagnostic plots for standardized and raw differences

  • Provides detailed convergence summary including:

    • Total nodes analyzed

    • Number and percentage of nodes converged at different thresholds

    • Maximum/minimum deviations

    • List of non-converged nodes (if any)

Examples

if (FALSE) { # \dontrun{
results <- mcmodule_converg(mc_results)
results <- mcmodule_converg(mc_results, from_quantile = 0.90, conv_threshold = 0.01)
} # }