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

Reinforcement learning design for ca...

Zhao, Yufan.

Linked to FindBook

Google Book

Amazon

博客來

Reinforcement learning design for cancer clinical trials.

Record Type:	Electronic resources : Monograph/item
Title/Author:	Reinforcement learning design for cancer clinical trials./
Author:	Zhao, Yufan.
Published:	Ann Arbor : ProQuest Dissertations & Theses, : 2009,
Description:	119 p.
Notes:	Source: Dissertation Abstracts International, Volume: 70-07, Section: B, page: 3862.
Contained By:	Dissertation Abstracts International70-07B.
Subject:	Biostatistics. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3366451
ISBN:	9781109277180

Reinforcement learning design for cancer clinical trials.
Zhao, Yufan.

Reinforcement learning design for cancer clinical trials. - Ann Arbor : ProQuest Dissertations & Theses, 2009 - 119 p.

Source: Dissertation Abstracts International, Volume: 70-07, Section: B, page: 3862.

Thesis (Ph.D.)--The University of North Carolina at Chapel Hill, 2009.

There has been significant recent research activity in developing therapies that are tailored to each individual. Finding such therapies in treatment settings involving multiple decision times is a major challenge. In this dissertation, we develop reinforcement learning trials for discovering these optimal regimens for life-threatening diseases such as cancer. A temporal-difference learning method called Q-learning is utilized which involves learning an optimal policy from a single training set of finite longitudinal patient trajectories. Approximating the Q-function with time-indexed parameters can be achieved by using support vector regression or extremely randomized trees. Within this framework, we demonstrate that the procedure can extract optimal strategies directly from clinical data without relying on the identification of any accurate mathematical models, unlike approaches based on adaptive design. We show that reinforcement learning has tremendous potential in clinical research because it can select actions that improve outcomes by taking into account delayed effects even when the relationship between actions and outcomes is not fully known.

ISBN: 9781109277180Subjects--Topical Terms:

1002712
Biostatistics.

Reinforcement learning design for cancer clinical trials.
LDR:03088nmm a2200325 4500 001 2164198
005 20181030085012.5
008 190424s2009 ||||||||||||||||| ||eng d
020 $a 9781109277180
035 $a (MiAaPQ)AAI3366451
035 $a (MiAaPQ)unc:10396
035 $a AAI3366451
040 $a MiAaPQ $c MiAaPQ
100 1 $a Zhao, Yufan. $3 3352243
245 1 0 $a Reinforcement learning design for cancer clinical trials.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2009
300 $a 119 p.
500 $a Source: Dissertation Abstracts International, Volume: 70-07, Section: B, page: 3862.
500 $a Adviser: Michael R. Kosorok.
502 $a Thesis (Ph.D.)--The University of North Carolina at Chapel Hill, 2009.
520 $a There has been significant recent research activity in developing therapies that are tailored to each individual. Finding such therapies in treatment settings involving multiple decision times is a major challenge. In this dissertation, we develop reinforcement learning trials for discovering these optimal regimens for life-threatening diseases such as cancer. A temporal-difference learning method called Q-learning is utilized which involves learning an optimal policy from a single training set of finite longitudinal patient trajectories. Approximating the Q-function with time-indexed parameters can be achieved by using support vector regression or extremely randomized trees. Within this framework, we demonstrate that the procedure can extract optimal strategies directly from clinical data without relying on the identification of any accurate mathematical models, unlike approaches based on adaptive design. We show that reinforcement learning has tremendous potential in clinical research because it can select actions that improve outcomes by taking into account delayed effects even when the relationship between actions and outcomes is not fully known.
520 $a To support our claims, the methodology's practical utility is firstly illustrated in a virtual simulated clinical trial. We then apply this general strategy with significant refinements to studying and discovering optimal treatments for advanced metastatic stage IIIB/IV non-small cell lung cancer (NSCLC). In addition to the complexity of the NSCLC problem of selecting optimal compounds for first and second-line treatments based on prognostic factors, another primary scientific goal is to determine the optimal time to initiate second-line therapy, either immediately or delayed after induction therapy, yielding the longest overall survival time. We show that reinforcement learning not only successfully identifies optimal strategies for two lines of treatment from clinical data, but also reliably selects the best initial time for second-line therapy while taking into account heterogeneities of NSCLC across patients.
590 $a School code: 0153.
650 4 $a Biostatistics. $3 1002712
650 4 $a Artificial intelligence. $3 516317
650 4 $a Statistics. $3 517247
690 $a 0308
690 $a 0800
690 $a 0463
710 2 $a The University of North Carolina at Chapel Hill. $b Biostatistics. $3 1023527
773 0 $t Dissertation Abstracts International $g 70-07B.
790 $a 0153
791 $a Ph.D.
792 $a 2009
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3366451