pagnani / PlmIsing

Pseudo likelihood maximization for pairwise Ising model

PseudoLikelihood Maximization for Ising Model

Documentation	Build Status	Coverage

The package compute the standard pseudolikelihood optimization on pairwise spin systems (Ising). The method exported is out=isingplmdca("nomefile",kwds...) analyzes data stored in nomefile file. Data should be a N x M formatted file of M configurations of N spins.

Returns out of type PlmOut with 3 fields:

1. out.pslike the pseudolikelihood (Vector{Float64})

2. out.J the coupling matrix

3. out.H the fields

Overview

The code uses NLopt which provides a Julia interfaces to the free/open-source NLopt library. The program can be run on multiple cores previous addprocs(nprocs) where nprocs should be some integer number np lower or equal to your (physical) number of cores.

Install

The package is not yet registered. To use it, clone the repository locally:

> git clone https://github.com/pagnani/PlmIsing.git

Then cd into the PlmIsing directory, and from the julia REPL do a

(PlmIsing) pkg> activate .
(PlmIsing) pkg> resolve

If you want to use it in parallel start julia with julia -p nprocs where nprocs is the number of process you want to use. Then from julia REPL

julia> @everywhere using Pkg
julia> @everywhere Pkg.activate(".")
julia> @everywhere using PlmIsing

Script

The inference can be also run with the plm_ising.jl script. It requires the installation of:

• ArgParse.jl:

julia> Pkg.add("ArgParse")

To use it, just run from the shell:

$ julia PATH-TO-PACKAGE/src/plm_ising.jl infile outfile

where infile is a file containing the Ising spin configurations (see above). Note that the script can be run in parallel just with

$ julia -p nproc PATH-TO-PACKAGE/src/plm_ising.jl infile outfile

where nproc is the (integer) number of cores.

Versions

Version 0.1.0 works with julia v0.6 Version 0.2.0 works with julia v1.0, v1.1, v1.2 Version 0.3.0 works with julia from v1.0 ... v1.5.*