View on GitHub

pomtc

Bayesian model-based clustering for longitudinal ordinal data

Bayesian model-based clustering for longitudinal ordinal data

Computer scripts to reproduce simulation in Costilla et al 2019

This repository contains the R and C++ binary files to reproduce results presented in Table 2 (section 3.4 Model validation using simulated data). C++ source files are also included for convenience. Scripts run in Linux, Windows and Mac OS X (x86-64, i.e. 64 bits versions).

A brief description of these files can be found below.

R scripts

C++ files

Binaries, x86-64 versions

Source files

Running instructions

To run the simulation please:

  1. Download or clone the git repository. Uncompress it to your local computer.
  2. Run “pomtc.sim.r” within R. Note that you need to access to several R libraries.
  3. Table 2 contents will be saved as a csv file.

For the simulation in the paper (n=1000, p=15, q=5, R=3) programs take about an hour to run using R 3.3.3 in a Xeon E5-2680 2.50GHz CPU. Depending on your computer specifications this time might vary. Running time includes simulating the data, estimating the model using 3 MCMC chains and relabeling. In addition to that, traceplots for the original and relabelled chains are also produced and the R session saved. The complete output is available here in R markdown format.

References

Costilla, Liu, Arnold, and Fernandez (2019). Bayesian model-based clustering for longitudinal ordinal data. Computational Statistics. https://doi.org/10.1007/s00180-019-00872-4

Comments/questions to

Roy Costilla

PhD Statistics

Institute for Molecular Biosciences and Queensland Alliance for Agriculture and Food Innovation. University of Queensland. Brisbane. Australia.

r.costilla@uq.edu.au

https://www.researchgate.net/profile/Roy_Costilla

https://twitter.com/CmRoycostilla