Preference Elicitation via Set-wise Max-margin

Implementation of the set-wise max-margin (setmargin for short) preference elicitation algorithm.

For details, see the paper:

Stefano Teso, Andrea Passerini, Paolo Viappiani. Constructive Preference Elicitation by Setwise Max-margin Learning. proceedings of 25th International Joint Conference on Artificial Intelligence. 2016. (To appear)

Requirements

This package requires the following packages:

• numpy
• gurobi
• matplotlib
• scikit-learn

Usage

To run the IJCAI-16 experiments, simply type:

    python ijcai16.py run-synthetic
    python ijcai16.py run-pc-with-costs

To perform preference elicitation on a specific dataset with given parameters, type:

    python ijcai16.py run $dataset $parameters

For instance, to run 20 trials of length 10 with set size 3 on the synthetic dataset, write:

    python ijcai16.py run synthetic -T 20 -n 10 -k 3

See:

    python ijcai16.py --help

for a full list of the accepted arguments.

Funding

The project is supported by the CARITRO Foundation through grant 2014.0372.