Parameter-free Online Linear Optimization with Side Information via Universal Coin Betting

Requirements

The experiments were run under the following environment:

python==3.8.1
matplotlib==3.3.4
numpy==1.20.0
pandas==1.2.1
py-cpuinfo==7.0.0
scikit-learn==0.24.1
scipy==1.6.0

Datasets

• The datasets ['BeijingPM2pt5', 'MetroInterstateTrafficVolume'] are included in the repository.

To replicate

• First, run exp-scripts.ipynb for each dataset. (This step can be skipped since the saved variables are also included in the repository for convenience.)
• Then, run generate-plots.ipynb to generate summary plots.