基于LSTM与动态行为模型在管控情景下预测Covid-19传播
物理与社会
2022-04-08 v1 机器学习
神经与进化计算
摘要
为准确预测Covid-19感染的区域传播,本研究提出了一种结合长短期记忆(LSTM)人工循环神经网络与动态行为模型的新型混合模型。若干因素和管控策略影响病毒传播,且Covid-19感染传播背后混杂变量所带来的不确定性十分显著。所提模型考虑了多因素的影响,以提升对十个受影响最严重国家及澳大利亚的病例数和死亡数的预测精度。结果表明,所提模型紧密复现了测试数据。它不仅提供准确预测,还估计了不确定性下系统的每日行为。该混合模型在可用数据有限的情况下优于LSTM模型。混合模型的参数针对每个国家使用遗传算法进行优化,以在考虑区域特性的同时提升预测能力。由于所提模型能在考量 containment 政策的情况下准确预测Covid-19传播,因此能够用于政策评估、规划与决策。
引用
@article{arxiv.2005.12270,
title = {Forecasting the Spread of Covid-19 Under Control Scenarios Using LSTM and Dynamic Behavioral Models},
author = {Seid Miad Zandavi and Taha Hossein Rashidi and Fatemeh Vafaee},
journal= {arXiv preprint arXiv:2005.12270},
year = {2022}
}
备注
As requested by the dear moderator, to assess the statistical significance of the reduction in RMSE in hybrid models compared to LSTM, each module was evaluated 500 times after hype-parameter tuning, and the corresponding RMSE distribution was used to estimate 95% confidence interval (CI) and t-test p-values comparing significant differences between different stages