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

On Asymptotic Distributions and Conf...

Zhao, Yu.

Linked to FindBook

Google Book

Amazon

博客來

On Asymptotic Distributions and Confidence Intervals for LIFT Measures in Data Mining.

Record Type:	Language materials, printed : Monograph/item
Title/Author:	On Asymptotic Distributions and Confidence Intervals for LIFT Measures in Data Mining./
Author:	Zhao, Yu.
Description:	82 p.
Notes:	Source: Dissertation Abstracts International, Volume: 75-07(E), Section: B.
Contained By:	Dissertation Abstracts International75-07B(E).
Subject:	Statistics. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3615547
ISBN:	9781303815676

On Asymptotic Distributions and Confidence Intervals for LIFT Measures in Data Mining.
Zhao, Yu.

On Asymptotic Distributions and Confidence Intervals for LIFT Measures in Data Mining. - 82 p.

Source: Dissertation Abstracts International, Volume: 75-07(E), Section: B.

Thesis (Ph.D.)--Northwestern University, 2014.

A LIFT measure, such as the response rate, lift, or the percentage of captured response, is a fundamental measure of effectiveness for a scoring rule obtained from data mining, which is estimated from a set of validation data. The LIFT measures are related to the ROC (Receiver Operator Characteristic), but there exist some important differences. In this paper, we study how to construct confidence intervals of the LIFT measures. We point out the difficulty of this task and how a simple binomial confidence interval for the response rate can miss variation from an estimated percentile of the scoring rule. We derive the asymptotic distribution using some advanced empirical process theory and the functional delta method in Appendix B. The additional variation is shown to be related to a conditional mean response, which can be estimated by a local averaging of the responses over the scores from the validation data. Alternatively, a subsampling method is shown to provide a valid confidence interval, without needing to estimate the conditional mean response. Numerical experiments are conducted to compare these different methods regarding the coverage probabilities and the lengths of the resulting confidence intervals.

ISBN: 9781303815676Subjects--Topical Terms:

517247
Statistics.

On Asymptotic Distributions and Confidence Intervals for LIFT Measures in Data Mining.
LDR:02120nam a2200277 4500 001 1969186
005 20141222143619.5
008 150210s2014 ||||||||||||||||| ||eng d
020 $a 9781303815676
035 $a (MiAaPQ)AAI3615547
035 $a AAI3615547
040 $a MiAaPQ $c MiAaPQ
100 1 $a Zhao, Yu. $3 1277443
245 1 0 $a On Asymptotic Distributions and Confidence Intervals for LIFT Measures in Data Mining.
300 $a 82 p.
500 $a Source: Dissertation Abstracts International, Volume: 75-07(E), Section: B.
500 $a Adviser: Wenxin Jiang.
502 $a Thesis (Ph.D.)--Northwestern University, 2014.
520 $a A LIFT measure, such as the response rate, lift, or the percentage of captured response, is a fundamental measure of effectiveness for a scoring rule obtained from data mining, which is estimated from a set of validation data. The LIFT measures are related to the ROC (Receiver Operator Characteristic), but there exist some important differences. In this paper, we study how to construct confidence intervals of the LIFT measures. We point out the difficulty of this task and how a simple binomial confidence interval for the response rate can miss variation from an estimated percentile of the scoring rule. We derive the asymptotic distribution using some advanced empirical process theory and the functional delta method in Appendix B. The additional variation is shown to be related to a conditional mean response, which can be estimated by a local averaging of the responses over the scores from the validation data. Alternatively, a subsampling method is shown to provide a valid confidence interval, without needing to estimate the conditional mean response. Numerical experiments are conducted to compare these different methods regarding the coverage probabilities and the lengths of the resulting confidence intervals.
590 $a School code: 0163.
650 4 $a Statistics. $3 517247
650 4 $a Business Administration, Marketing. $3 1017573
690 $a 0463
690 $a 0338
710 2 $a Northwestern University. $b Statistics. $3 1673808
773 0 $t Dissertation Abstracts International $g 75-07B(E).
790 $a 0163
791 $a Ph.D.
792 $a 2014
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3615547