Skip to content
Code for UAI paper « Sampling and Inference for Beta Neutral-to-the-Left Models of Sparse Networks »
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
examples
.gitignore
LICENSE
README.md
REQUIRE
crp.jl
dataset.jl
evaluation.jl
experiments_util.jl
geometric_ia.jl
gibbs_util.jl
likelihoods.jl
ntl_gibbs.jl
poisson_ia.jl
slice.jl
synthetic.jl

README.md

NTL.jl

Code to accompany UAI 2018 paper 'Sampling and Inference for Beta Neutral-to-the-Left Models of Sparse Networks' (supplement).

Data

SNAP datasets

The Stanford Large Network Dataset Collection contains a number of interesting temporal networks. We recommend preprocessing the datasets as follows:

wget https://snap.stanford.edu/data/$NAME.txt.gz
gunzip $NAME.txt.gz
sort -k3 -n $NAME.txt > sorted-$NAME.txt

Maximum likelihood parameter estimation

See examples/mle.jl for an example of computing MLEs on massive datasets.

Gibbs sampling arrival order, arrival times, parameters

See examples/gibbs.jl for an example of performing posterior inference over parameters and latent variables on datasets of modest size (e.g., hundreds or thousands of nodes). The code can be run interactively (i.e., section by section) or as a script from the Julia REPL or command line.

See examples/gibbs_plots.jl for some example plots for assessing sampler output.

See examples/gibbs_ess_experiments.jl for code used to produce the tables in Section 5.1 of the paper.

You can’t perform that action at this time.