analiza aiqbench
This study empirically evaluates forecasting architectures—including world models, Temporal Fusion Transformers, and ensemble averages—across multiple retail data grains to determine optimal model selection for supply chain planning. The findings indicate that no single architecture dominates universally; instead, performance varies significantly based on the decision granularity, with simple averaging often outperforming complex deep learning approaches in specific hierarchical contexts.
World Models, Trees and Foundation Models for Retail Forecasting: Evidence from Four Public Data Sets
arXiv:2610.25170v1 Announce Type: new Abstract: A retailer makes different decisions at different grains of the product, store and time hierarchy: store orders for an item in a store, supplier funding for an item across stores, category performance by store, and the financial plan for the chain. Which forecasting model to use depends on the grain. This paper compares models at each grain on four public data sets (M5, six Dominick's categories, Breakfast at the Frat, and Corporaci\'on Favorita, the last kept out of all development), over rolling forecast origins with paired tests and model confidence sets. It covers price and promotion through the plan inputs these data record, not product, assortment or place. The models are a world model that carries each item's state forward under the price and promotion plan, the same network without that state, LightGBM, the Temporal Fusion Transformer, TimesFM, Chronos-Bolt, Timestral, seasonal naive, and an equal-weight average of the world model, LightGBM and TimesFM given the plan. The answer changed with the grain. For an item by store and an item across stores, the average had the lowest error on seven and eight of nine data sets, and the 90% confidence set held a median of 1 and 2 models. For a category by store, a category and the chain total, the lowest-error model differed between data sets and a median of 4 of nine models were not separated by the four to six windows available. For category-by-store planning cost, the average or LightGBM was lowest on seven of eight data sets, and the world model was lowest only on M5; it was behind LightGBM item by store on six of eight. Every trained model used the plan, and on the weekly promotion-driven data pretrained models were less accurate without it. These results hold for these data sets; a forecast solved across all four Ps, or a retailer's own data, may give different results. Every model, test and change that lost is reported.
Abstract:A retailer makes different decisions at different grains of the product, store and time hierarchy: store orders for an item in a store, supplier funding for an item across stores, category performance by store, and the financial plan for the chain. Which forecasting model to use depends on the grain. This paper compares models at each grain on four public data sets (M5, six Dominick's categories, Breakfast at the Frat, and Corporación Favorita, the last kept out of all development), over rolling forecast origins with paired tests and model confidence sets. It covers price and promotion through the plan inputs these data record, not product, assortment or place. The models are a world model that carries each item's state forward under the price and promotion plan, the same network without that state, LightGBM, the Temporal Fusion Transformer, TimesFM, Chronos-Bolt, Timestral, seasonal naive, and an equal-weight average of the world model, LightGBM and TimesFM given the plan. The answer changed with the grain. For an item by store and an item across stores, the average had the lowest error on seven and eight of nine data sets, and the 90% confidence set held a median of 1 and 2 models. For a category by store, a category and the chain total, the lowest-error model differed between data sets and a median of 4 of nine models were not separated by the four to six windows available. For category-by-store planning cost, the average or LightGBM was lowest on seven of eight data sets, and the world model was lowest only on M5; it was behind LightGBM item by store on six of eight. Every trained model used the plan, and on the weekly promotion-driven data pretrained models were less accurate without it. These results hold for these data sets; a forecast solved across all four Ps, or a retailer's own data, may give different results. Every model, test and change that lost is reported.
37 pages, 18 figures, 26 tables. Code: this https URL, archived at doi:https://doi.org/10.5281/zenodo.23448056
Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
arXiv:2610.25170 [cs.LG]
(or arXiv:2610.25170v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2610.25170 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 430 cuvinte
Subiecte conexe
Prophet vs NeuralProphet vs TimeGPT vs Chronos: A Practical Comparison
Analytics Vidhya · 9 iun. 2026
Expert-Guided Forecast Editing for Time-Series Foundation Models
arXiv · 23 iul. 2026
GlucoFM-Bench: Benchmarking Time-Series Foundation Models for Blood Glucose Forecasting
arXiv · 8 iun. 2026
Beyond Numerical Time Series: A Unified Benchmark for Multimodal Forecasting with Heterogeneous Context
arXiv · 16 sept. 2026