MCMCLib  

MCMCLib is a lightweight C++ library of Markov Chain Monte Carlo (MCMC) methods.

Features:

• Parallelized C++11 implementations of several well-known MCMC methods, including:

  • Random Walk Metropolis-Hastings (RWMH);
  • Metropolis-adjusted Langevin algorithm (MALA);
  • Hamiltonian Monte Carlo (HMC); and
  • Riemannian Manifold HMC.

• Samplers designed for multi-modal distributions:

  • Equi-Energy sampling; and
  • Differential Evolution (DE).

• Built on the Armadillo C++ linear algebra library for fast and efficient matrix-based computation.

• Released under a permissive, non-GPL license.

Status

The library is actively maintained, and is still being extended.

Algorithms:

• rwmh
• mala
• hmc
• rmhmc
• aees
• de

Syntax

MCMCLib functions are generally defined as

algorithm(<initial values>, <draws output>, <log kernel target distribution>, <optional: data for target distribution>, <optional: algorithm settings>)

where the inputs, in order, are:

• a vector of initial values that define the starting point for the algorithm, and will contain the solution vector at completion;
• the objective function to be minimized (or zeroed-out);
• (optional) any additional parameters passed to the objective function; and
• (optional) control and tuning parameters for the MCMC algorithms.

For example, the RWMH algorithm is called using:

bool rwmh(const arma::vec& initial_vals, arma::mat& draws_out, std::function<double (const arma::vec& vals_inp, void* target_data)> target_log_kernel, void* target_data);

Installation

The library can be installed on Unix-alike systems via the standard ./configure && make method:

# clone MCMCLib into the current directory
git clone https://github.com/kthohr/mcmc ./mcmc
# build and install
cd ./mcmc
./configure -i "/usr/local" -p
make
make install

The last line will install MCMCLib to /usr/local.

There are several configure options available (./configure -h):

• -c a coverage build (used with Codecov)
• -d a 'development' build
• -g a debugging build (optimization flags set to -O0 -g)
• -h print help
• -i install path; default: current directory.
• -m specify the BLAS and Lapack libraries to link against; for example, -m "-lopenblas" or -m "-framework Accelerate"
• -o compiler optimization options; defaults to -O3 -march=native -ffp-contract=fast -flto -DARMA_NO_DEBUG
• -p enable OpenMP parallelization features (recommended)

Armadillo

MCMCLib is built on the Armadillo C++ linear algebra library. The build script will search for Armadillo files in the usual places: /usr/include, /usr/local/include, /opt/include, /opt/local/include. If the Armadillo header files were installed to a different location, set:

export ARMA_INCLUDE_PATH=/path/to/armadillo

before running ./configure. Otherwise the build script will download the required files from the Armadillo GitHub repository.

Example

Objective: Sample the mean parameter from a normal distribution.

Code:

#include "mcmc.hpp"

struct norm_data {
    double sigma;
    arma::vec x;

    double mu_0;
    double sigma_0;
};

double ll_dens(const arma::vec& vals_inp, void* ll_data)
{
    const double mu = vals_inp(0);
    const double pi = arma::datum::pi;

    norm_data* dta = reinterpret_cast<norm_data*>(ll_data);
    const double sigma = dta->sigma;
    const arma::vec x = dta->x;

    const int n_vals = x.n_rows;

    //

    const double ret = - ((double) n_vals) * (0.5*std::log(2*pi) + std::log(sigma)) - arma::accu( arma::pow(x - mu,2) / (2*sigma*sigma) );

    //

    return ret;
}

double log_pr_dens(const arma::vec& vals_inp, void* ll_data)
{
    norm_data* dta = reinterpret_cast< norm_data* >(ll_data);

    const double mu_0 = dta->mu_0;
    const double sigma_0 = dta->sigma_0;
    const double pi = arma::datum::pi;

    const double x = vals_inp(0);

    const double ret = - 0.5*std::log(2*pi) - std::log(sigma_0) - std::pow(x - mu_0,2) / (2*sigma_0*sigma_0);

    return ret;
}

double log_target_dens(const arma::vec& vals_inp, void* ll_data)
{
    return ll_dens(vals_inp,ll_data) + log_pr_dens(vals_inp,ll_data);
}

int main()
{
    const int n_data = 100; // simulated data length
    const double mu = 2.0;  // true mean

    norm_data dta;
    dta.sigma = 1.0;
    dta.mu_0 = 1.0;
    dta.sigma_0 = 2.0;

    arma::vec x_dta = mu + arma::randn(n_data,1);
    dta.x = x_dta;

    arma::vec initial_val(1);
    initial_val(0) = 1.0;

    arma::mat draws_out;
    mcmc::rwmh(initial_val,draws_out,log_target_dens,&dta);

    return 0;
}

See http://www.kthohr.com/mcmclib.html for a detailed description of each algorithm, and more examples.

Author

Keith O'Hara

License

Apache Version 2