Modeling Reference-dependent Choices with Graph Neural Networks

TL;DR

This paper tackles modeling reference-dependent consumer choices within recommender systems by introducing ArcRec, a deep learning framework that builds an attribute-level Reference Network from co-purchase data and decomposes it into Attributed Reference Networks (ARNs) for fine-grained reference points. It learns a reference-dependent utility $r_{u,i}$ that combines both interest-inspired and price-inspired preferences, weighted by an Attribute-level Willingness-To-Pay (AWTP) that captures attribute salience, and trains end-to-end via Bayesian Personalized Ranking. ArcRec demonstrates superior performance over fourteen baselines on synthetic and real-world data, including cold-start scenarios, and provides interpretable insights through AWTP analyses and consumer segmentation. The approach advances practical recommendation by integrating reference effects, price interactions, and attribute-level trade-offs, enabling more personalized and price-aware recommendations with improved interpretability. The results have meaningful implications for pricing strategies, feature prioritization, and targeted marketing in e-commerce.

Abstract

While the classic Prospect Theory has highlighted the reference-dependent and comparative nature of consumers' product evaluation processes, few models have successfully integrated this theoretical hypothesis into data-driven preference quantification, particularly in the realm of recommender systems development. To bridge this gap, we propose a new research problem of modeling reference-dependent preferences from a data-driven perspective, and design a novel deep learning-based framework named Attributed Reference-dependent Choice Model for Recommendation (ArcRec) to tackle the inherent challenges associated with this problem. ArcRec features in building a reference network from aggregated historical purchase records for instantiating theoretical reference points, which is then decomposed into product attribute specific sub-networks and represented through Graph Neural Networks. In this way, the reference points of a consumer can be encoded at the attribute-level individually from her past experiences but also reflect the crowd influences. ArcRec also makes novel contributions to quantifying consumers' reference-dependent preferences using a deep neural network-based utility function that integrates both interest-inspired and price-inspired preferences, with their complex interaction effects captured by an attribute-aware price sensitivity mechanism. Most importantly, ArcRec introduces a novel Attribute-level Willingness-To-Pay measure to the reference-dependent utility function, which captures a consumer's heterogeneous salience of product attributes via observing her attribute-level price tolerance to a product. Empirical evaluations on both synthetic and real-world online shopping datasets demonstrate ArcRec's superior performances over fourteen state-of-the-art baselines.

Figures (6)

• Figure 1: Example of Reference Network and Attributed Reference Network. (a) users' movie watching history; (b) reference network constructed from the movie watching history; (c) genre-specific reference network; (d) director-specific reference network.
• Figure 2: The architecture of ArcRec.
• Figure 3: Ranking performances on synthetic data.
• Figure 4: Visualization of consumers' attribute-level willingness to pay (AWTP). (a) consumer segmentation based on AWTP. (b) Heat map of sampled consumers' AWTP.
• Figure 5: Representative consumers' AWTP and purchasing records.

Theorems & Definitions (2)

• Definition 3.1: Reference Network
• Definition 3.2: Attributed Reference Network