Cross-channel Recommendation for Multi-channel Retail

Yijin Choi,Jongkyung Shin,Chiehyeon Lim

2024-07-18

Abstract:An increasing number of retailers are expanding their channels to the offline and online domains, transforming them into multi-channel retailers. This transition emphasizes the need for cross-channel recommendations. Given that each retail channel represents a separate domain with a unique context, this can be regarded as a cross-domain recommendation (CDR). However, existing studies on CDR did not address the scenarios where both users and items partially overlap across multi-retail channels which we define as "cross-channel retail recommendation (CCRR)". This paper introduces our original work on CCRR using a real-world dataset from a multi-channel retail store. Specifically, we study significant challenges in integrating user preferences across both channels and propose a novel model for CCRR using a channel-wise attention mechanism. We empirically validate our model's superiority in addressing CCRR over existing models. Finally, we offer implications for future research on CCRR, delving into our experiment results.

Information Retrieval

What problem does this paper attempt to address?

This paper attempts to address the problem of how to perform Cross-Channel Retail Recommendation (CCRR) in a multi-channel retail environment. Specifically, the paper focuses on how to integrate user preferences from different channels (online and offline) when there is partial overlap between users and items, in order to improve the performance of the recommendation system.

### Main Challenges
1. **Partial Overlap of Users and Items**: Existing Cross-Domain Recommendation (CDR) research usually considers the overlap of either users or items, but does not consider the case where both are partially overlapping.
2. **Channel-Specific and Channel-Shared Preferences**: Users' purchasing behaviors may differ across different channels, so it is necessary to distinguish between channel-specific preferences and channel-shared preferences.
3. **Complexity of Cross-Channel Integration**: Interactions in different channels occur in different contexts, making cross-channel recommendation more complex.

### Solution
To address these challenges, the authors propose a new model called C2Rec. C2Rec addresses the above issues through the following mechanisms:

1. **Channel Attention Mechanism**: Introduces a channel attention mechanism to distinguish user preferences in each channel and weight preferences according to different channels.
2. **Joint Learning**: Simultaneously performs recommendation tasks and interaction classification tasks to better capture user behavior characteristics in different channels.
3. **Auxiliary Objective**: Designs an auxiliary objective to adjust attention scores, ensuring that attention scores reflect users' purchasing behaviors in different channels.

### Experimental Validation
The authors conducted experiments on a real-world multi-channel retail dataset, and the results show that C2Rec outperforms existing Single-Domain Recommendation (SDR) and Cross-Domain Recommendation (CDR) baseline models on multiple evaluation metrics.

### Conclusion
This study defines the CCRR scenario for the first time and proposes an effective model to address this problem. Through the channel attention mechanism and joint learning, C2Rec can better integrate user preferences across different channels, thereby improving the performance of the recommendation system. This research provides valuable insights for future multi-channel and omni-channel retailers' recommendation systems.