基于带外生变量的广义网络自回归模型预测采购经理人指数以量化经济对COVID-19管控与死亡率的响应
统计方法学
2021-07-19 v1 应用统计
摘要
了解经济当前状态、其如何响应COVID-19管控与指标，以及未来可能走势十分重要。我们使用近期发展的广义网络自回归（GNAR）模型，基于贸易决定网络，对多个国家的采购经理人指数（Purchasing Managers' Indices）进行建模与预测。我们所用的网络将国家相连，其中连边本身或其权重由各国间出口贸易程度决定。我们将这些模型扩展以纳入节点特定时间序列外生变量（GNARX模型），借此将COVID-19管控严格指数与COVID-19死亡率纳入分析。高度简洁的GNAR模型在均方预测误差上大幅优于向量自回归模型，而我们的GNARX模型本身优于GNAR模型。进一步的混频建模预测了英国经济受更严、更弱或无干预影响的程度。
引用
@article{arxiv.2107.07605,
title = {Quantifying the economic response to COVID-19 mitigations and death rates via forecasting Purchasing Managers' Indices using Generalised Network Autoregressive models with exogenous variables},
author = {Guy P Nason and James L Wei},
journal= {arXiv preprint arXiv:2107.07605},
year = {2021}
}
备注
To be read before the Royal Statistical Society at the Society's 2021 annual conference held in Manchester on Wednesday, September 8th 2021, the President, Professor Sylvia Richardson, in the Chair. Accepted by the Journal of the Royal Statistical Society, Series A