TY - UNPB
T1 - New Item Consumption Prediction Using Deep Learning
AU - Shekasta, Michael
AU - Katz, Gilad
AU - Greenstein-Messica, Asnat
AU - Rokach, Lior
AU - Shapira, Bracha
PY - 2019/5/5
Y1 - 2019/5/5
N2 - Recommendation systems have become ubiquitous in today's online world and are an integral part of practically every e-commerce platform. While traditional recommender systems use customer history, this approach is not feasible in 'cold start' scenarios. Such scenarios include the need to produce recommendations for new or unregistered users and the introduction of new items. In this study, we present the Purchase Intent Session-bAsed (PISA) algorithm, a content-based algorithm for predicting the purchase intent for cold start session-based scenarios. Our approach employs deep learning techniques both for modeling the content and purchase intent prediction. Our experiments show that PISA outperforms a well-known deep learning baseline when new items are introduced. In addition, while content-based approaches often fail to perform well in highly imbalanced datasets, our approach successfully handles such cases. Finally, our experiments show that combining PISA with the baseline in non-cold start scenarios further improves performance.
AB - Recommendation systems have become ubiquitous in today's online world and are an integral part of practically every e-commerce platform. While traditional recommender systems use customer history, this approach is not feasible in 'cold start' scenarios. Such scenarios include the need to produce recommendations for new or unregistered users and the introduction of new items. In this study, we present the Purchase Intent Session-bAsed (PISA) algorithm, a content-based algorithm for predicting the purchase intent for cold start session-based scenarios. Our approach employs deep learning techniques both for modeling the content and purchase intent prediction. Our experiments show that PISA outperforms a well-known deep learning baseline when new items are introduced. In addition, while content-based approaches often fail to perform well in highly imbalanced datasets, our approach successfully handles such cases. Finally, our experiments show that combining PISA with the baseline in non-cold start scenarios further improves performance.
KW - cs.LG
KW - cs.IR
U2 - 10.48550/arXiv.1905.01686
DO - 10.48550/arXiv.1905.01686
M3 - Preprint
BT - New Item Consumption Prediction Using Deep Learning
ER -