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

Supporting joint human-computer judg...

University of Illinois at Urbana-Champaign.

Linked to FindBook

Google Book

Amazon

博客來

Supporting joint human-computer judgment under uncertainty.

Record Type:	Language materials, printed : Monograph/item
Title/Author:	Supporting joint human-computer judgment under uncertainty./
Author:	Miller, Sarah.
Description:	135 p.
Notes:	Adviser: Alex Kirlik.
Contained By:	Dissertation Abstracts International70-02B.
Subject:	Engineering, Industrial. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3347452
ISBN:	9781109025170

Supporting joint human-computer judgment under uncertainty.
Miller, Sarah.

Supporting joint human-computer judgment under uncertainty. - 135 p.

Adviser: Alex Kirlik.

Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 2008.

Recent high-profile failures of computational models to make accurate predictions, estimations or forecasts (e.g., risk models used on Wall Street) highlight the importance of keeping humans in the loop in a variety operational contexts, often involving judgment and decision making. The purpose of this dissertation was to design and evaluate a joint, human-computer architecture that integrates those aspects of a judgment task that can be best supported by computational models and those aspects best left to human expertise. Computational techniques are typically based on judgment or decision models that may integrate information more accurately and reliably than unaided human cognition. Yet humans may have knowledge transcending what can be formally modeled, either in principle, or because of access to information the model does not have.

ISBN: 9781109025170Subjects--Topical Terms:

626639
Engineering, Industrial.

Supporting joint human-computer judgment under uncertainty.
LDR:03011nam 2200301 a 45 001 855640
005 20100708
008 100708s2008 ||||||||||||||||| ||eng d
020 $a 9781109025170
035 $a (UMI)AAI3347452
035 $a AAI3347452
040 $a UMI $c UMI
100 1 $a Miller, Sarah. $3 1022286
245 1 0 $a Supporting joint human-computer judgment under uncertainty.
300 $a 135 p.
500 $a Adviser: Alex Kirlik.
500 $a Source: Dissertation Abstracts International, Volume: 70-02, Section: B, page: 1268.
502 $a Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 2008.
520 $a Recent high-profile failures of computational models to make accurate predictions, estimations or forecasts (e.g., risk models used on Wall Street) highlight the importance of keeping humans in the loop in a variety operational contexts, often involving judgment and decision making. The purpose of this dissertation was to design and evaluate a joint, human-computer architecture that integrates those aspects of a judgment task that can be best supported by computational models and those aspects best left to human expertise. Computational techniques are typically based on judgment or decision models that may integrate information more accurately and reliably than unaided human cognition. Yet humans may have knowledge transcending what can be formally modeled, either in principle, or because of access to information the model does not have.
520 $a Three experiments were conducted to evaluate and refine the architecture design. The experimental tasks were patterned after the popular games of fantasy baseball and fantasy football, which require competitors, near the beginning of each season, to predict the performance of professional baseball or football players over the course of that season on the basis of past performance and other relevant information. Some aspects of this predictive judgment may be best performed computationally (by linear regression models) and other aspects may be best performed by human expertise. The most substantial and challenging aspects of interface design involved how to welcome expert input on a case-by-case (player-by-player) basis, yet also provide visual guidance for, and constraint on, how these case-specific inputs should also reflect an appropriate degree of regression to the mean, or reliance on base-rate, rather than case-specific, information.
520 $a Overall, the results revealed that the integrated, human-computer architecture supported including humans into the loop without loss in judgment quality. In some cases, the joint, human-computer system outperformed either the human or the model alone. Implications and future directions are discussed.
590 $a School code: 0090.
650 4 $a Engineering, Industrial. $3 626639
650 4 $a Psychology, Experimental. $3 517106
690 $a 0546
690 $a 0623
710 2 $a University of Illinois at Urbana-Champaign. $3 626646
773 0 $t Dissertation Abstracts International $g 70-02B.
790 $a 0090
790 1 0 $a Kirlik, Alex, $e advisor
791 $a Ph.D.
792 $a 2008
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3347452