A Unified Contextual Bandit Framework for Long- and Short-Term Recommendations
Abstract
We present a unified contextual bandit framework for recommendation problems that is able to capture long- and short-term interests of users. The model is devised in dual space and the derivation is consequentially carried out using Fenchel-Legrende conjugates and thus leverages to a wide range of tasks and settings. We detail two instantiations for regression and classification scenarios and obtain well-known algorithms for these special cases. The resulting general and unified framework allows for quickly adapting contextual bandits to different applications at-hand. The empirical study demonstrates that the proposed long- and short-term framework outperforms both, short-term and long-term models on data. Moreover, a tweak of the combined model proves beneficial in cold start problems.
Cite
Text
Tavakol and Brefeld. "A Unified Contextual Bandit Framework for Long- and Short-Term Recommendations." European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, 2017. doi:10.1007/978-3-319-71246-8_17Markdown
[Tavakol and Brefeld. "A Unified Contextual Bandit Framework for Long- and Short-Term Recommendations." European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, 2017.](https://mlanthology.org/ecmlpkdd/2017/tavakol2017ecmlpkdd-unified/) doi:10.1007/978-3-319-71246-8_17BibTeX
@inproceedings{tavakol2017ecmlpkdd-unified,
title = {{A Unified Contextual Bandit Framework for Long- and Short-Term Recommendations}},
author = {Tavakol, Maryam and Brefeld, Ulf},
booktitle = {European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases},
year = {2017},
pages = {269-284},
doi = {10.1007/978-3-319-71246-8_17},
url = {https://mlanthology.org/ecmlpkdd/2017/tavakol2017ecmlpkdd-unified/}
}