東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

FindBook

Google Book

Amazon

博客來

Understanding and Prediction of Dynamical Systems with Machine Learning.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Understanding and Prediction of Dynamical Systems with Machine Learning./
作者:	Hou, Xiao.
出版者:	Ann Arbor : ProQuest Dissertations & Theses, : 2021,
面頁冊數:	142 p.
附註:	Source: Dissertations Abstracts International, Volume: 83-02, Section: B.
Contained By:	Dissertations Abstracts International83-02B.
標題:	Applied mathematics. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=28718956
ISBN:	9798538106479

Understanding and Prediction of Dynamical Systems with Machine Learning.
Hou, Xiao.

Understanding and Prediction of Dynamical Systems with Machine Learning. - Ann Arbor : ProQuest Dissertations & Theses, 2021 - 142 p.

Source: Dissertations Abstracts International, Volume: 83-02, Section: B.

Thesis (Ph.D.)--The University of Wisconsin - Madison, 2021.

This item must not be sold to any third party vendors.

Complex nonlinear dynamical systems are frequently utilized and play a significant role in the study of natural science and many other fields. This thesis studies dynamical systems both theoretically and numerically with the help of machine learning, quantifying uncertainties in the system and developing models and algorithms to explore two specific application problems. An information criterion is first introduced to quantify model error of dynamical systems, and then two specific applications of machine learning algorithms on the study of dynamical systems are presented in the context of COVID-19 pandemic modeling and partial differential equation (PDE) solvers, respectively.The first part of this thesis introduces a simple information criterion that is developed to quantify the model error in imperfect models, incorporating not only the equilibrium statistics but also temporal information of the system. Three applications of this information criterion are presented, including the study of model reduction, stochastic parameterizations and intermittent events, all of which are illustrated with detailed examples. The second part of this thesis investigates a specific example of the combination of machine learning algorithms and dynamical systems in the context of COVID-19 pandemic in 2020. A new human mobility flow-augmented stochastic SEIR model is developed, which employs machine learning community detection method, data assimilation technique, and online learning of parameters with a stochastic parameterization approach. Numerical simulations are performed to reconstruct the historical transmission trajectories of COVID-19 in the two largest counties in Wisconsin, and the results are further analyzed to inspect the associations between COVID-19 transmission and human mobility as well as demographic features. Several scenario studies are performed during the study, providing insights for policymakers on the design of effective regionalization-based policies to control the ongoing spread of COVID-19. The reliability of community detection results and the validity of the dynamical model are further inspected by conducting numerous sensitivity tests and building several new models for comparison.As the last work of this thesis, a new numerical algorithm is proposed and analyzed in order to solve hyperbolic conservation laws. This newly designed numerical scheme combines the Lax Wendroff method with $l_1$ regularization and, comparing to past work with similar techniques, adds a new critical conservation constraint. It is proved that this scheme is consistent, convergent, total variation diminishing, and satisfies the weak entropy condition for conservation laws. Numerical solutions to Burgers and Euler's equation are presented to validate the analytical results and to explore possible improvement to the algorithm.

ISBN: 9798538106479Subjects--Topical Terms:

2122814
Applied mathematics.
Subjects--Index Terms:

Dynamical systems

Understanding and Prediction of Dynamical Systems with Machine Learning.
LDR:03921nmm a2200325 4500 001 2343893
005 20220513114345.5
008 241004s2021 ||||||||||||||||| ||eng d
020 $a 9798538106479
035 $a (MiAaPQ)AAI28718956
035 $a AAI28718956
040 $a MiAaPQ $c MiAaPQ
100 1 $a Hou, Xiao. $3 3682568
245 1 0 $a Understanding and Prediction of Dynamical Systems with Machine Learning.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2021
300 $a 142 p.
500 $a Source: Dissertations Abstracts International, Volume: 83-02, Section: B.
500 $a Advisor: Li, Qin;Chen, Nan.
502 $a Thesis (Ph.D.)--The University of Wisconsin - Madison, 2021.
506 $a This item must not be sold to any third party vendors.
520 $a Complex nonlinear dynamical systems are frequently utilized and play a significant role in the study of natural science and many other fields. This thesis studies dynamical systems both theoretically and numerically with the help of machine learning, quantifying uncertainties in the system and developing models and algorithms to explore two specific application problems. An information criterion is first introduced to quantify model error of dynamical systems, and then two specific applications of machine learning algorithms on the study of dynamical systems are presented in the context of COVID-19 pandemic modeling and partial differential equation (PDE) solvers, respectively.The first part of this thesis introduces a simple information criterion that is developed to quantify the model error in imperfect models, incorporating not only the equilibrium statistics but also temporal information of the system. Three applications of this information criterion are presented, including the study of model reduction, stochastic parameterizations and intermittent events, all of which are illustrated with detailed examples. The second part of this thesis investigates a specific example of the combination of machine learning algorithms and dynamical systems in the context of COVID-19 pandemic in 2020. A new human mobility flow-augmented stochastic SEIR model is developed, which employs machine learning community detection method, data assimilation technique, and online learning of parameters with a stochastic parameterization approach. Numerical simulations are performed to reconstruct the historical transmission trajectories of COVID-19 in the two largest counties in Wisconsin, and the results are further analyzed to inspect the associations between COVID-19 transmission and human mobility as well as demographic features. Several scenario studies are performed during the study, providing insights for policymakers on the design of effective regionalization-based policies to control the ongoing spread of COVID-19. The reliability of community detection results and the validity of the dynamical model are further inspected by conducting numerous sensitivity tests and building several new models for comparison.As the last work of this thesis, a new numerical algorithm is proposed and analyzed in order to solve hyperbolic conservation laws. This newly designed numerical scheme combines the Lax Wendroff method with $l_1$ regularization and, comparing to past work with similar techniques, adds a new critical conservation constraint. It is proved that this scheme is consistent, convergent, total variation diminishing, and satisfies the weak entropy condition for conservation laws. Numerical solutions to Burgers and Euler's equation are presented to validate the analytical results and to explore possible improvement to the algorithm.
590 $a School code: 0262.
650 4 $a Applied mathematics. $3 2122814
650 4 $a Mathematics. $3 515831
650 4 $a Infections. $3 1621997
650 4 $a Conservation laws. $3 3682569
650 4 $a Partial differential equations. $3 2180177
650 4 $a Age. $3 1486010
650 4 $a Pandemics. $3 3464080
650 4 $a Equilibrium. $3 668417
650 4 $a Data assimilation. $3 3682570
650 4 $a Population density. $3 908253
650 4 $a Traffic flow. $3 756182
650 4 $a Stochastic models. $3 764002
650 4 $a Algorithms. $3 536374
650 4 $a Ethnicity. $3 519288
650 4 $a Kalman filters. $3 3682571
650 4 $a Dynamical systems. $3 3219613
650 4 $a Coronaviruses. $3 894828
650 4 $a Distance learning. $3 3557921
650 4 $a Mobility. $3 3557724
650 4 $a Disease transmission. $3 3560843
650 4 $a COVID-19. $3 3554449
653 $a Dynamical systems
653 $a Machine learning
690 $a 0364
690 $a 0405
710 2 $a The University of Wisconsin - Madison. $b Mathematics. $3 2101076
773 0 $t Dissertations Abstracts International $g 83-02B.
790 $a 0262
791 $a Ph.D.
792 $a 2021
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=28718956