We describe the solution of team ISMLL for the ECML-PKDD 2016 Discovery Challenge on Bank Card Usage for both tasks. Our solution is based on three pillars. Gradient boosted decision trees as a strong regression and classification model, an intensive search for good hyperparameter configurations and strong features that exploit geolocation information. This approach achieved the best performance on the public leaderboard for the first task and a decent fourth position for the second task.
Cite
@article{arxiv.1610.03996,
title = {Bank Card Usage Prediction Exploiting Geolocation Information},
author = {Martin Wistuba and Nghia Duong-Trung and Nicolas Schilling and Lars Schmidt-Thieme},
journal= {arXiv preprint arXiv:1610.03996},
year = {2016}
}
Comments
Describes the winning solution for the ECML-PKDD 2016 Discovery Challenge on Bank Card Usage Analysis. Final results on the private leaderboard are available here: https://dms.sztaki.hu/ecml-pkkd-2016/#/app/privateleaderboard