A Unified Contextual Bandit Framework for Long- and Short-Term Recommendations
Publikation: Beiträge in Sammelwerken › Aufsätze in Konferenzbänden › Forschung › begutachtet
Standard
A Unified Contextual Bandit Framework for Long- and Short-Term Recommendations. / Tavakol, Maryam; Brefeld, Ulf.
Machine Learning and Knowledge Discovery in Databases: European Conference, ECML PKDD 2017 Skopje, Macedonia, September 18 – 22, 2017; Proceedings, Part II. Hrsg. / Michelangelo Ceci; Jaakko Hollmen; Ljupco Todorovski; Celine Vens; Saso Dzeroski. Band 2 Springer, 2017. S. 269-284 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Band 10535 LNAI).Publikation: Beiträge in Sammelwerken › Aufsätze in Konferenzbänden › Forschung › begutachtet
Harvard
APA
Vancouver
Bibtex
}
RIS
TY - CHAP
T1 - A Unified Contextual Bandit Framework for Long- and Short-Term Recommendations
AU - Tavakol, Maryam
AU - Brefeld, Ulf
N1 - Conference code: 27
PY - 2017/12/30
Y1 - 2017/12/30
N2 - 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.
AB - 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.
KW - Informatics
KW - Machine learning
KW - Recommender systems
KW - Contextual bandits
KW - Recommendation
KW - Dual optimization
KW - Personalization
UR - http://ecmlpkdd2017.ijs.si/papers/paperID239.pdf
UR - http://www.scopus.com/inward/record.url?scp=85040222425&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-71246-8_17
DO - 10.1007/978-3-319-71246-8_17
M3 - Article in conference proceedings
SN - 978-3-319-71245-1
VL - 2
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 269
EP - 284
BT - Machine Learning and Knowledge Discovery in Databases
A2 - Ceci, Michelangelo
A2 - Hollmen, Jaakko
A2 - Todorovski, Ljupco
A2 - Vens, Celine
A2 - Dzeroski, Saso
PB - Springer
T2 - THE EUROPEAN CONFERENCE ON MACHINE LEARNING & PRINCIPLES AND PRACTICE OF KNOWLEDGE DISCOVERY IN DATABASES 2017
Y2 - 18 September 2017 through 22 September 2017
ER -