Multiple Wallet Estimation for E-commerce Marketplaces

Yehuda Kayam, M.Sc. student at the department of Industrial

2 June 2022, 12:30 PM, Room 206& via zoom

Abstract:
Wallet estimation considers the prediction of customers spending in the retail industry. Previous work showed that customer wallets may be modelled by high quantile regression. In this work, we generalize the wallet estimation scheme and consider a multi-wallet problem, where multiple wallets are estimated simultaneously. We argue that a naive approach, which considers every wallet separately, can be improved since customers spending on different items are typically correlated. We introduce a multi-target learning (MTL) framework, based on deep-neural networks, and apply it to large scale e-commerce vendors. We show that the MTL based method significantly improves the prediction accuracy, compared to currently known schemes.
 

Bio:
Yehuda Kayam, M.Sc. student at the Department of Industrial Engineering and management in Tel Aviv university, specializing in business analytics. Yehuda holds a B.Sc degree in Industrial Engineering from the Technion. His research is supervised by Dr. Amichai Painski