SAFERec: Self-Attention and Frequency Enriched Model for Next Basket Recommendation

TL;DR

SAFERec is a new model that combines self-attention with item frequency data to improve next-basket recommendation, outperforming existing methods especially in recall metrics.

Contribution

It introduces SAFERec, a novel approach that enhances transformer models with frequency information for better next-basket recommendations.

Findings

SAFERec outperforms baseline models in multiple datasets.

Achieves an 8% improvement in Recall@10.

Integrates frequency data into transformer architecture effectively.

Abstract

Transformer-based approaches such as BERT4Rec and SASRec demonstrate strong performance in Next Item Recommendation (NIR) tasks. However, applying these architectures to Next-Basket Recommendation (NBR) tasks, which often involve highly repetitive interactions, is challenging due to the vast number of possible item combinations in a basket. Moreover, frequency-based methods such as TIFU-KNN and UP-CF still demonstrate strong performance in NBR tasks, frequently outperforming deep-learning approaches. This paper introduces SAFERec, a novel algorithm for NBR that enhances transformer-based architectures from NIR by incorporating item frequency information, consequently improving their applicability to NBR tasks. Extensive experiments on multiple datasets show that SAFERec outperforms all other baselines, specifically achieving an 8\% improvement in Recall@10.