東華大學圖書館 |

Language: English

Help

回圖書館首頁

手機版館藏查詢

Back

Switch To: Labeled | MARC Mode | ISBD

Linked to FindBook

Google Book

Amazon

博客來

Robust Statistical Methods for Model Selection and Land Cover Change Monitoring.

Record Type:	Electronic resources : Monograph/item
Title/Author:	Robust Statistical Methods for Model Selection and Land Cover Change Monitoring./
Author:	Wendelberger, Laura Jean.
Description:	1 online resource (131 pages)
Notes:	Source: Dissertations Abstracts International, Volume: 84-04, Section: B.
Contained By:	Dissertations Abstracts International84-04B.
Subject:	Remote sensing. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=29342802click for full text (PQDT)
ISBN:	9798352601945

Robust Statistical Methods for Model Selection and Land Cover Change Monitoring.
Wendelberger, Laura Jean.

Robust Statistical Methods for Model Selection and Land Cover Change Monitoring. - 1 online resource (131 pages)

Source: Dissertations Abstracts International, Volume: 84-04, Section: B.

Thesis (Ph.D.)--North Carolina State University, 2022.

Includes bibliographical references

Interpretable statistical models are valuable collaborative tools that make incorporating expert knowledge, gleaning insights from analyses, and iterating on hypotheses easier. Robust model selection should not be sensitive to a particular sample, and studying model ensembles instead of a single model mitigates the sometimes restrictive form of an interpretable model while acknowledging uncertainty about the exact form of the model. Bayesian methods are a natural setting to incorporate model uncertainty because they weight model contributions based on the data. In this dissertation, we explore robust methods for model selection and change detection using penalized regression and exact Bayesian models to acknowledge model uncertainty.In the first project, we consider model uncertainty in the context of model selection, which often aims to choose a single model, assuming that the form of the model is correct. However, there may be multiple possible underlying explanatory patterns in a set of predictors that could explain a response. Model selection without regard for model uncertainty can fail to bring these patterns to light. We explore multi-model penalized regression (MMPR) to acknowledge model uncertainty in the context of penalized regression. We examine how different penalty settings can promote either shrinkage or sparsity of coefficients in separate models. The method is tuned to explicitly limit model similarity. A choice of penalty form that enforces variable selection is applied to predict stacking fault energy (SFE) from steel alloy composition. The aim is to identify multiple models with different subsets of covariates that explain a single type of response.In the second and third projects, we consider near real time change detection in the context of Earth monitoring applications using remote sensing data. We introduce Robust Online Bayesian Monitoring (roboBayes), which extends Bayesian Online Changepoint Detection (BOCPD; Adams and MacKay, 2007) to be robust against occasional outliers without compromising the computational efficiency of an exact posterior change distribution nor the detection latency. Without this robustness feature, the change detection algorithm cannot be employed autonomously without an inflated false positive rate. We show via simulations that the method effectively detects change in the presence of outliers. The method is then applied to monitor deforestation in Myanmar where we show competitive performance compared to current online changepoint detection methods with fewer limitations on when it can be applied.In the third project, we propose MultiResolution roboBayes (MR roboBayes) to incorporate spatial information into image monitoring. Instead of analyzing pixel intensities, we propose representing images using multiresolution, spatially localized basis functions calculated by a Discrete Wavelet Transform (DWT; Mallat, 1989). We posit several ways to recombine change information from wavelet space into real space. MR roboBayes supplies an opportunity for computational speedup while still incorporating information from each pixel in different resolutions. We show via simulation that redundancy of detection in multiple components is useful for differentiating true from false detections. We demonstrate MR roboBayes for heavy construction detection in two different regions.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2023

Mode of access: World Wide Web

ISBN: 9798352601945Subjects--Topical Terms:

535394
Remote sensing.
Index Terms--Genre/Form:

542853
Electronic books.

Robust Statistical Methods for Model Selection and Land Cover Change Monitoring.
LDR:04619nmm a2200325K 4500 001 2354642
005 20230428105637.5
006 m o d
007 cr mn ---uuuuu
008 241011s2022 xx obm 000 0 eng d
020 $a 9798352601945
035 $a (MiAaPQ)AAI29342802
035 $a (MiAaPQ)NCState_Univ18402039733
035 $a AAI29342802
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Wendelberger, Laura Jean. $3 3694999
245 1 0 $a Robust Statistical Methods for Model Selection and Land Cover Change Monitoring.
264 0 $c 2022
300 $a 1 online resource (131 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 84-04, Section: B.
500 $a Advisor: Reich, Brian J.; Wilson, Alyson G.
502 $a Thesis (Ph.D.)--North Carolina State University, 2022.
504 $a Includes bibliographical references
520 $a Interpretable statistical models are valuable collaborative tools that make incorporating expert knowledge, gleaning insights from analyses, and iterating on hypotheses easier. Robust model selection should not be sensitive to a particular sample, and studying model ensembles instead of a single model mitigates the sometimes restrictive form of an interpretable model while acknowledging uncertainty about the exact form of the model. Bayesian methods are a natural setting to incorporate model uncertainty because they weight model contributions based on the data. In this dissertation, we explore robust methods for model selection and change detection using penalized regression and exact Bayesian models to acknowledge model uncertainty.In the first project, we consider model uncertainty in the context of model selection, which often aims to choose a single model, assuming that the form of the model is correct. However, there may be multiple possible underlying explanatory patterns in a set of predictors that could explain a response. Model selection without regard for model uncertainty can fail to bring these patterns to light. We explore multi-model penalized regression (MMPR) to acknowledge model uncertainty in the context of penalized regression. We examine how different penalty settings can promote either shrinkage or sparsity of coefficients in separate models. The method is tuned to explicitly limit model similarity. A choice of penalty form that enforces variable selection is applied to predict stacking fault energy (SFE) from steel alloy composition. The aim is to identify multiple models with different subsets of covariates that explain a single type of response.In the second and third projects, we consider near real time change detection in the context of Earth monitoring applications using remote sensing data. We introduce Robust Online Bayesian Monitoring (roboBayes), which extends Bayesian Online Changepoint Detection (BOCPD; Adams and MacKay, 2007) to be robust against occasional outliers without compromising the computational efficiency of an exact posterior change distribution nor the detection latency. Without this robustness feature, the change detection algorithm cannot be employed autonomously without an inflated false positive rate. We show via simulations that the method effectively detects change in the presence of outliers. The method is then applied to monitor deforestation in Myanmar where we show competitive performance compared to current online changepoint detection methods with fewer limitations on when it can be applied.In the third project, we propose MultiResolution roboBayes (MR roboBayes) to incorporate spatial information into image monitoring. Instead of analyzing pixel intensities, we propose representing images using multiresolution, spatially localized basis functions calculated by a Discrete Wavelet Transform (DWT; Mallat, 1989). We posit several ways to recombine change information from wavelet space into real space. MR roboBayes supplies an opportunity for computational speedup while still incorporating information from each pixel in different resolutions. We show via simulation that redundancy of detection in multiple components is useful for differentiating true from false detections. We demonstrate MR roboBayes for heavy construction detection in two different regions.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2023
538 $a Mode of access: World Wide Web
650 4 $a Remote sensing. $3 535394
650 4 $a Funding. $3 3557916
650 4 $a Deforestation. $3 600701
655 7 $a Electronic books. $2 lcsh $3 542853
690 $a 0799
710 2 $a ProQuest Information and Learning Co. $3 783688
710 2 $a North Carolina State University. $3 1018772
773 0 $t Dissertations Abstracts International $g 84-04B.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=29342802 $z click for full text (PQDT)