FreshRetailNet-LT: A Stockout-Annotated Censored Demand Dataset for Latent Demand Recovery and Forecasting in Fresh Retail
arXiv:2505.16319v3 Announce Type: replace Abstract: Accurate demand estimation is critical for the retail business in guiding the inventory and pricing policies of perishable products. However, it faces fundamental challenges from censored sales data during stockouts, where unobserved demand creates systemic policy biases. Existing datasets lack the temporal resolution and annotations needed to address this censoring effect. To fill this gap, we present FreshRetailNet-50K, the first large-scale benchmark for censored demand estimation. It comprises 50,000 store-product time series of detailed
The increasing complexity of supply chains and the maturity of AI/ML techniques for data analysis make it possible to address long-standing challenges in retail demand forecasting.
Accurate demand forecasting, especially for perishable goods and during stockouts, directly impacts profitability, waste reduction, and inventory management for retailers.
Retailers now have access to a specialized, large-scale benchmark for developing more sophisticated AI models to handle censored demand, leading to improved operational efficiency.
- · Retailers of perishable goods
- · AI/ML researchers and data scientists
- · Supply chain software developers
- · Consumers (reduced stockouts)
- · Traditional inventory management systems
- · Retailers slow to adopt advanced AI
Improved demand forecasting reduces food waste and optimizes inventory levels in fresh retail.
More efficient retail operations lead to better pricing strategies and increased profit margins, potentially impacting consumer prices.
The success in fresh retail could spur the development of similar specialized datasets and AI solutions for other complex supply chains, accelerating automation in logistics.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG