EventCast: Hybrid Demand Forecasting in E-Commerce with LLM-Based Event Knowledge

2026-02-07

AI Summary

This study addresses the challenge of demand forecasting in e-commerce during high-impact events such as major promotions and holidays, where traditional methods often fail due to abrupt demand fluctuations. The authors propose EventCast, a novel framework that leverages large language models (LLMs) not for direct numerical prediction but for event reasoning—extracting future event signals from unstructured business data and generating interpretable textual summaries enriched with cultural常识. These summaries are fused with historical time-series features via a dual-tower architecture to produce accurate and explainable forecasts. Evaluated across 160 regions in four countries over ten months of real-world operations, EventCast reduces MAE and MSE by up to 86.9% and 97.7%, respectively, compared to models without event knowledge, and achieves 57.0% and 83.3% improvements over industrial baselines during event periods.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Demand forecasting is a cornerstone of e-commerce operations, directly impacting inventory planning and fulfillment scheduling. However, existing forecasting systems often fail during high-impact periods such as flash sales, holiday campaigns, and sudden policy interventions, where demand patterns shift abruptly and unpredictably. In this paper, we introduce EventCast, a modular forecasting framework that integrates future event knowledge into time-series prediction. Unlike prior approaches that ignore future interventions or directly use large language models (LLMs) for numerical forecasting, EventCast leverages LLMs solely for event-driven reasoning. Unstructured business data, which covers campaigns, holiday schedules, and seller incentives, from existing operational databases, is processed by an LLM that converts it into interpretable textual summaries leveraging world knowledge for cultural nuances and novel event combinations. These summaries are fused with historical demand features within a dual-tower architecture, enabling accurate, explainable, and scalable forecasts. Deployed on real-world e-commerce scenarios spanning 4 countries of 160 regions over 10 months, EventCast achieves up to 86.9% and 97.7% improvement on MAE and MSE compared to the variant without event knowledge, and reduces MAE by up to 57.0% and MSE by 83.3% versus the best industrial baseline during event-driven periods. EventCast has deployed into real-world industrial pipelines since March 2025, offering a practical solution for improving operational decision-making in dynamic e-commerce environments.

Problem

• demand forecasting
• e-commerce
• event-driven demand
• time-series prediction
• high-impact events

Innovation

• LLM-based event reasoning
• hybrid demand forecasting
• dual-tower architecture
• event-driven time series
• e-commerce operations