Change time estimation uncertainty in nonlinear dynamical systems with applications to COVID‐19

Change time estimation uncertainty in nonlinear dynamical systems with applications to COVID‐19
复制标题

改变非线性动力系统中的时间估计不确定性及其在 COVID-19 中的应用

DOI:
10.1002/rnc.5974
复制
发表时间:
2022
影响因子:
3.9
通讯作者:
Sandberg, Henrik
Sandberg, Henrik
中科院分区:
计算机科学3区
文献类型:
--
作者:
Alisic, Rijad;Paré, Philip E.;Sandberg, Henrik

文献摘要

参考文献

相似文献

由于大多数国家快速连续实施了几项非药物干预措施,因此很难估计每项非药物干预措施(NPI)对COVID-19传播速度的影响。在这篇文章中,我们分析了在测量噪声的存在下,非线性动力系统的参数突然变化的可检测性,该参数可用于表示NPI或病毒的突变。具体来说,通过采取不可知的方法,我们提供了必要的条件时,最好的可能无偏估计是能够隔离的影响,突然改变模型参数,通过使用Hammersley-Chapman-Robbins(HCR)下界。给出了HCR下界计算的几种简化方法,这些方法依赖于突变的幅值和系统的动力学特性。我们进一步定义了基于两个输出轨迹之间的最大距离的信息量最大的样本的概念,这是HCR下限收敛的一个很好的指标。这些结果随后用于分析易感-感染-移除模型。例如,我们表明,使用恢复/死亡的数量进行分析,而不是感染的累积数量,可能是一个较差的信号,因为突然的变化从根本上更难以估计,似乎需要更多的样本。最后,通过模拟验证了这些结果,并将其应用于COVID-19在法国传播的真实的数据。
The impact that each individual non‐pharmaceutical intervention (NPI) had on the spread rate of COVID‐19 is difficult to estimate, since several NPIs were implemented in rapid succession in most countries. In this article, we analyze the detectability of sudden changes in a parameter of nonlinear dynamical systems, which could be used to represent NPIs or mutations of the virus, in the presence of measurement noise. Specifically, by taking an agnostic approach, we provide necessary conditions for when the best possible unbiased estimator is able to isolate the effect of a sudden change in a model parameter, by using the Hammersley–Chapman–Robbins (HCR) lower bound. Several simplifications to the calculation of the HCR lower bound are given, which depend on the amplitude of the sudden change and the dynamics of the system. We further define the concept of the most informative sample based on the largestdistance between two output trajectories, which is a good indicator of when the HCR lower bound converges. These results are thereafter used to analyze the susceptible‐infected‐removed model. For instance, we show that performing analysis using the number of recovered/deceased, as opposed to the cumulative number of infected, may be an inferior signal to use since sudden changes are fundamentally more difficult to estimate and seem to require more samples. Finally, these results are verified by simulations and applied to real data from the spread of COVID‐19 in France.
DOI: --
发表时间: 1961
期刊:
影响因子: --
作者:
Guy Lebanon
通讯作者: Guy Lebanon
墨西哥城的三级护理中心的严重共同19岁的住院死亡率;死亡原因,危险因素和医院饱和的影响。
DOI: 10.1371/journal.pone.0245772
发表时间: 2021
期刊: PloS one
影响因子: 3.7
作者:
Olivas-Martínez A;Cárdenas-Fragoso JL;Jiménez JV;Lozano-Cruz OA;Ortiz-Brizuela E;Tovar-Méndez VH;Medrano-Borromeo C;Martínez-Valenzuela A;Román-Montes CM;Martínez-Guerra B;González-Lara MF;Hernandez-Gilsoul T;Herrero AG;Tamez-Flores KM;Ochoa-Hein E;Ponce-de-León A;Galindo-Fraga A;Kershenobich-Stalnikowitz D;Sifuentes-Osornio J
通讯作者: Sifuentes-Osornio J
动态系统中的隐私
DOI: 10.1007/978-981-15-0493-8
发表时间: 2020
期刊: 2019 American Control Conference (ACC)
影响因子: --
作者:
K. Teranishi;M. Kusaka;Naoki Shimada;J. Ueda;K. Kogiso
通讯作者: K. Kogiso
Hammersley-Chapman-Robbins 脉冲响应数据的极点和残差估计下限
DOI: 10.1109/globalsip.2017.8308657
发表时间: 2017
期刊: IEEE GlobalSIP
影响因子: --
作者:
Maruf, Abdullah Al;Roy, Sandip
通讯作者: Roy, Sandip
信息论中的某些不等式和克拉默-饶不等式
DOI: --
发表时间: 1954
期刊:
影响因子: --
作者:
S. Kullback
通讯作者: S. Kullback