Replication code repository accompanying the paper On Sinkhorn's Algorithm and Choice Modeling.

The code for replicating Table 2 in Appendix E.1 can be found in the notebooks folder. Each row of Table 1 corresponds to a distinct dataset, and can be replicated by running the corresponding Jupyter notebook.

The code for replicating Figure 1 in Appendix E.2 can be found in the simulations folder. Figure 1 can be generated by running getplot.m.

The NASCAR 2002 dataset is available through the paper Hunter (2004). The SUSHI dataset on Sushi preferences is provided by Toshihiro Kamishima. The GIFGIF dataset is provided by Lucas Maystre.