experiments/README.md

In edahelsinki/cyclesampler: Randomization Algorithms for Large Sparse Matrices

Experiments

The files in this folder are used to run the experiments for the following paper:

Kai Puolamäki, Andreas Henelius, Antti Ukkonen. Randomization algorithms for large sparse networks. Physical Review E 99, 053311, 2019. https://doi.org/10.1103/PhysRevE.99.053311

Obtaining the data

The directory 'data' contains bash-scripts and an R-script that can be used to download the datasets. Each of the bash-scripts are run as, e.g., ./get_movielens.sh /tmp/movielens where the only argument is to some suitable temporary directory. The R-script is run as, e.g., Rscript --vanilla get_itn_data.R /tmp/itndata.

There are also Python (version 2) scripts and an R-script ("preprocess_datasets.R") for preprocessing some of the datasets.

Examples and Experiments

The examples are contained in the files example_1.R and example_2.R.

The experiments on bipartite recommender datasets are contained in the following files:

• convergence.R Sample networks where the vertex strengths are preserved exactly, while the edge weights w(e) are allowed to vary subject to interval constraints.
• convergence_nwfix.R Sample networks where both edge weights and vertex strengths are allowed to vary subject to interval constraints.
• analyse_results.R Used to produce the convergence and runtime experiments, after first having run convergence.R and convergence_nwfix.R.

The experiments on the International Trade Network (ITN) data is contained in the following files:

• analyse_itn.R Calculate the average clustering coefficient for the ITN networks.
• analyse_itn_results.R Produce plots with confidence bands for the average clustering coefficient for surrogates for the ITN data, after first having run analyse_itn.R.

edahelsinki/cyclesampler documentation built on June 9, 2019, 10:51 a.m.