F-FOMAML: GNN-Enhanced Meta-Learning for Peak Period Demand Forecasting with Proxy Data

TL;DR

This paper introduces F-FOMAML, a GNN-enhanced meta-learning approach that leverages proxy data and relational metadata to improve demand forecasting during peak periods, addressing data scarcity challenges.

Contribution

It proposes a novel GNN-based meta-learning algorithm that uses proxy data and domain-specific metadata to enhance peak demand prediction accuracy.

Findings

Achieves 26.24% MAE reduction on vending machine data

Improves demand forecast accuracy on JD.com dataset

Theoretically guarantees better generalization with more training tasks

Abstract

Demand prediction is a crucial task for e-commerce and physical retail businesses, especially during high-stake sales events. However, the limited availability of historical data from these peak periods poses a significant challenge for traditional forecasting methods. In this paper, we propose a novel approach that leverages strategically chosen proxy data reflective of potential sales patterns from similar entities during non-peak periods, enriched by features learned from a graph neural networks (GNNs)-based forecasting model, to predict demand during peak events. We formulate the demand prediction as a meta-learning problem and develop the Feature-based First-Order Model-Agnostic Meta-Learning (F-FOMAML) algorithm that leverages proxy data from non-peak periods and GNN-generated relational metadata to learn feature-specific layer parameters, thereby adapting to demand forecasts for peak events. Theoretically, we show that by considering domain similarities through task-specific metadata, our model achieves improved generalization, where the excess risk decreases as the number of training tasks increases. Empirical evaluations on large-scale industrial datasets demonstrate the superiority of our approach. Compared to existing state-of-the-art models, our method demonstrates a notable improvement in demand prediction accuracy, reducing the Mean Absolute Error by 26.24% on an internal vending machine dataset and by 1.04% on the publicly accessible JD.com dataset.