Fit a Modified Connor-Mosimann Distribution to Elicited Quantiles in Multinomial Problems

Fits a modified version of the Connor-Mosimann distribution (Connor & Mosimann (1969)), a Connor-Mosimann distribution or Dirichlet distribution (e.g. Gelman, Carlin, Stern & Rubin Chapter 3.5 (2004, ) to elicited quantiles of a multinomial distribution. Code is also provided to sample from the distributions, generating inputs suitable for a probabilistic sensitivity analysis / Monte Carlo simulation in a decision model.


The purpose of this package is to fit a modified Connor-Mosimann distribution to quantiles for multinomial problems elicited from experts. It also includes functions to fit Connor-Mosimann and Dirichlet distributions, and to sample from modified Connor-Mosimann and Connor-Mosimann distributions for use in decision analytic models (rtools already provides a function to sample from a Dirichlet distribution).

For details on how to use please see the accompanying vignette.


Outcomes <- c("Remission","Progression","Dead")
RawData <- matrix(data = c(0.43, 0.55, 0.65,
                         0.16, 0.27, 0.46,
                         0.03, 0.18, 0.23
SearchParams <- c(10000,100) #number of iterations, max number of searches
ModCMorCM <- 1 # if 1 will fit mCM, if 0 will fit CM
Quantiles <- c(0.025,0.5,0.975) # example here is 95% credibility limits and median.
mCM <- fitModCM(Outcomes, RawData, SearchParams, ModCMorCM, Quantiles)
# Sample from the fitted distribution
n <- 100
Z <- mCM[1:(nrow(mCM)-1),1:4]
mCMSamples <- rModCM(n,Z)
colnames(mCMSamples) <- Outcomes

Installation instructions



modcmfitr 0.1.0

  • First release

Reference manual

It appears you don't have a PDF plugin for this browser. You can click here to download the reference manual.


0.1.0 by Ed Wilson, 4 years ago

Browse source code at

Authors: Ed Wilson [aut, cre]

Documentation:   PDF Manual  

GPL-2 license

Imports gtools, nloptr

Suggests cowplot, ggplot2, tidyr, knitr, rmarkdown

See at CRAN