Research on Hierarchical Early Warning of Drug Shortages Based on Probability Prediction and Interval Estimation

Read full text Summarise with VIRAT
Authors: Zhonghua Pang, Yuhan Feng, Tingyuan Jiang, Fuhao Jing, Pengzhan Tang, Yali Zhao
Cited by: 0
References: 5

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Abstract Objective To improve medicine shortage/demand forecasting and early warning during public-health shocks. Methods Weekly drug-demand series were built from routine consumption and inventory data. VMD was used for multi-scale decomposition, NRBO optimized LSSVM parameters, and DeepAR–LSTM modelled nonlinear patterns and uncertainty. Models were evaluated via rolling-origin backtesting and compared with ARIMA. Warnings triggered when predicted demand exceeded inventory plus safety stock. Results The hybrid model demonstrated lower errors and stronger robustness than ARIMA, with clearer advantages during shock periods. Conclusion The framework enables 7–30-day forecasts with prediction intervals, and provides technical support for hierarchical early warning and procurement decision-making of drug shortages.

Fields of Study & Subjects

Forecasting Techniques and Applications Supply Chain and Inventory Management Stock Market Forecasting Methods

Publication details

Journal / source
Research Square
Publication date
26 March 2026
Language
EN
Document type
Preprint