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

Efficient Dynamic Prediction of High...

Jin, Ying,

FindBook

Google Book

Amazon

博客來

Efficient Dynamic Prediction of High-Density Generalized Data With Complex Structure and the Evaluation of Predictive Performance /

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Efficient Dynamic Prediction of High-Density Generalized Data With Complex Structure and the Evaluation of Predictive Performance // Ying Jin.
作者:	Jin, Ying,
面頁冊數:	1 electronic resource (125 pages)
附註:	Source: Dissertations Abstracts International, Volume: 86-06, Section: B.
Contained By:	Dissertations Abstracts International86-06B.
標題:	Biostatistics. -
電子資源:	https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31636423
ISBN:	9798346858515

Efficient Dynamic Prediction of High-Density Generalized Data With Complex Structure and the Evaluation of Predictive Performance /
Jin, Ying,

Efficient Dynamic Prediction of High-Density Generalized Data With Complex Structure and the Evaluation of Predictive Performance /Ying Jin. - 1 electronic resource (125 pages)

Source: Dissertations Abstracts International, Volume: 86-06, Section: B.

Dynamic prediction, which typically refers to the prediction of future outcomes using historical records, is a major topic of statistical research. In biomedical studies, generalized outcomes are often of interest, such as survival outcomes, biomarkers, physical activity records, etc. These outcomes are often time-dependent, thus the prediction would update with the observation history. Developed under this overarching theme, this dissertation aims to address the gap in current literature on the evaluation of predictive performance and develop novel efficient methods for large-scale, high-density data.The first project is a comparison study of different measures of discriminative performance of time-to-event models, in particular, time-dependent AUC and concordance. Many estimators for these quantities have been proposed, which can be broadly categorized as either semiparametric estimators or non-parametric estimators. This project reviews various estimators' mathematical constructions and compares the behavior of the two classes of estimators. Importantly, we identify a previously unknown feature of the class of semi-parametric estimators that can result in vastly overoptimistic out-of-sample estimation of discriminative performance in common applied tasks, suggesting that they may be inappropriate for use in model assessment and selection based on out-of-sample evaluation criteria. Non-parametric estimators, which do not exhibit this behavior, are highly unstable for local discrimination, though the issue can be mitigated through penalized regression splines smoothing.The second project proposes a novel method for dynamic prediction of high-density generalized functional data, in practice often collected as dense repeated measures with complex correlation structure. This method is developed based on a recently developed approach estimating complex patterns of variation for exponential family data: fast Generalized Functional Principal Components Analysis (fGFPCA). It accommodates a highly flexible model structure and can make accurate predictions without reestimating parameters, which is a notable extension of traditional methods. Its advantages in efficiency, flexibility, and predictive performance are illustrated through a simulation study and a data application on the minute-level activity indicator from National Health and Nutrition Examination Survey (NHANES) 2011-2014 accelerometer data. In the third project, we extend this method to multilevel functional data, where more than one series of measures are observed along a dense grid. We also explore a new application of these method to the application of missing values.

English

ISBN: 9798346858515Subjects--Topical Terms:

1002712
Biostatistics.
Subjects--Index Terms:

Computational efficiency

Efficient Dynamic Prediction of High-Density Generalized Data With Complex Structure and the Evaluation of Predictive Performance /
LDR:04291nmm a22004573i 4500 001 2429417
005 20250430100823.5
006 m o d
007 cr|nu||||||||
008 261011s2024 miu||||||m |||||||eng d
020 $a 9798346858515
035 $a (MiAaPQD)AAI31636423
035 $a AAI31636423
040 $a MiAaPQD $b eng $c MiAaPQD $e rda
100 1 $a Jin, Ying, $e author. $3 3817254
245 1 0 $a Efficient Dynamic Prediction of High-Density Generalized Data With Complex Structure and the Evaluation of Predictive Performance / $c Ying Jin.
264 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2024
300 $a 1 electronic resource (125 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: 86-06, Section: B.
500 $a Advisors: Leroux, Andrew Committee members: Juarez-Colunga, Elizabeth; Wrobel, Julia; Wang, Yue; Shapiro, Allison L.
502 $b Ph.D. $c University of Colorado Denver, Anschutz Medical Campus $d 2024.
520 $a Dynamic prediction, which typically refers to the prediction of future outcomes using historical records, is a major topic of statistical research. In biomedical studies, generalized outcomes are often of interest, such as survival outcomes, biomarkers, physical activity records, etc. These outcomes are often time-dependent, thus the prediction would update with the observation history. Developed under this overarching theme, this dissertation aims to address the gap in current literature on the evaluation of predictive performance and develop novel efficient methods for large-scale, high-density data.The first project is a comparison study of different measures of discriminative performance of time-to-event models, in particular, time-dependent AUC and concordance. Many estimators for these quantities have been proposed, which can be broadly categorized as either semiparametric estimators or non-parametric estimators. This project reviews various estimators' mathematical constructions and compares the behavior of the two classes of estimators. Importantly, we identify a previously unknown feature of the class of semi-parametric estimators that can result in vastly overoptimistic out-of-sample estimation of discriminative performance in common applied tasks, suggesting that they may be inappropriate for use in model assessment and selection based on out-of-sample evaluation criteria. Non-parametric estimators, which do not exhibit this behavior, are highly unstable for local discrimination, though the issue can be mitigated through penalized regression splines smoothing.The second project proposes a novel method for dynamic prediction of high-density generalized functional data, in practice often collected as dense repeated measures with complex correlation structure. This method is developed based on a recently developed approach estimating complex patterns of variation for exponential family data: fast Generalized Functional Principal Components Analysis (fGFPCA). It accommodates a highly flexible model structure and can make accurate predictions without reestimating parameters, which is a notable extension of traditional methods. Its advantages in efficiency, flexibility, and predictive performance are illustrated through a simulation study and a data application on the minute-level activity indicator from National Health and Nutrition Examination Survey (NHANES) 2011-2014 accelerometer data. In the third project, we extend this method to multilevel functional data, where more than one series of measures are observed along a dense grid. We also explore a new application of these method to the application of missing values.
546 $a English
590 $a School code: 1639
650 4 $a Biostatistics. $3 1002712
650 4 $a Health sciences. $3 3168359
650 4 $a Computer science. $3 523869
650 4 $a Bioinformatics. $3 553671
653 $a Computational efficiency
653 $a Concordance
653 $a Dynamic prediction
653 $a Functional data
653 $a Generalized repeated measures
653 $a Time-dependent AUC
690 $a 0308
690 $a 0566
690 $a 0984
690 $a 0715
710 2 $a University of Colorado Denver, Anschutz Medical Campus. $b Biostatistics. $e degree granting institution. $3 3817255
720 1 $a Leroux, Andrew $e degree supervisor.
773 0 $t Dissertations Abstracts International $g 86-06B.
790 $a 1639
791 $a Ph.D.
792 $a 2024
856 4 0 $u https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31636423