Next-basket prediction in a high-dimensional setting using gated recurrent units

作者：

Highlights：

• Product-basket purchase patterns can be accurately learned from data.

• Our GRU-based next basket prediction model outperforms a state-of-the-art benchmark.

• Our proposed model is scalable to assortments with thousands of products.

• Inclusion of basket-specific covariates leads to (minor) performance improvements.

摘要

•Product-basket purchase patterns can be accurately learned from data.•Our GRU-based next basket prediction model outperforms a state-of-the-art benchmark.•Our proposed model is scalable to assortments with thousands of products.•Inclusion of basket-specific covariates leads to (minor) performance improvements.

论文关键词：Next-basket prediction,Gated recurrent units,Neural networks,Machine learning

论文评审过程：Received 9 May 2022, Revised 5 September 2022, Accepted 5 September 2022, Available online 10 September 2022, Version of Record 19 September 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.118795