A knowledge-based framework that turns demand forecasts into risk-aware ordering decisions on the public M5 retail dataset. Gradient-boosted trees and an LSTM are compared with two baselines; SHAP explains the forecasts; forecast-error risk is calibrated on validation data and frozen for test; and a rule-based quantile order policy lowers held-out decision cost versus ordering the point forecast. Includes all experiment code, figures, tables and logs.
No takes yet. Share an insight, caveat, or question.
Niketh Sai Avirneni (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: