In this paper, we detailedly describe our solution for the IEEE BigData Cup 2021: RL-based RecSys (Track 1: Item Combination Prediction). We first conduct an exploratory data analysis on the dataset and then utilize the findings to design our framework. Specifically, we use a two-headed transformer-based network to predict user feedback and unlocked sessions, along with the proposed session-aware reweighted loss, multi-tasking with click behavior prediction, and randomness-in-session augmentation. In the final private leaderboard on Kaggle, our method ranked 2nd with a categorization accuracy of 0.39224.
@article{arxiv.2111.07154,
title = {Session-aware Item-combination Recommendation with Transformer Network},
author = {Tzu-Heng Lin and Chen Gao},
journal= {arXiv preprint arXiv:2111.07154},
year = {2021}
}
Comments
2nd place solution in IEEE Bigdata Cup 2021 (Track 1: Item Combination Prediction). Our code is available at https://github.com/lzhbrian/bigdatacup2021