Evaluation of data aging: A technique for discounting old data during student modeling

Geoffrey I. Webb; Mark Kuzmycz

Evaluation of data aging: A technique for discounting old data during student modeling

Geoffrey I. Webb
, Mark Kuzmycz

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › Research › peer-review

Abstract

Student modeling systems must operate in an environment in which a student’s mastery of a subject matter is likely to change as a lesson progresses. A student model is formed from evaluation of evidence about the student’s mastery of the domain. However, given that such mastery will change, older evidence is likely to be less valuable than recent evidence. Data aging addresses this issue by discounting the value of older evidence. This paper provides experimental evaluation of the effects of data aging. While it is demonstrated that data aging can result in statistically significant increases in both the number and accuracy of predictions that a modeling system makes, it is also demonstrated that the reverse can be true. Further, the effects experienced are of only small magnitude. It is argued that these results demonstrate some potential for data aging as a general strategy, but do not warrant employing data aging in its current form.

Original language	English
Title of host publication	Intelligent Tutoring Systems - 4th International Conference, ITS 1998, Proceedings
Editors	Carol L. Redfield, Barry P. Goettl, Valerie J. Shute, Henry M. Halff
Publisher	Springer
Pages	384-393
Number of pages	10
ISBN (Print)	3540647708, 9783540647706
Publication status	Published - 1 Jan 1998
Externally published	Yes
Event	International Conference on Intelligent Tutoring Systems 1998 - San Antonio, United States of America Duration: 16 Aug 1998 → 19 Aug 1998 Conference number: 4th

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	1452
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	International Conference on Intelligent Tutoring Systems 1998
Abbreviated title	ITS 1998
Country/Territory	United States of America
City	San Antonio
Period	16/08/98 → 19/08/98

Cite this

Webb, G. I., & Kuzmycz, M. (1998). Evaluation of data aging: A technique for discounting old data during student modeling. In C. L. Redfield, B. P. Goettl, V. J. Shute, & H. M. Halff (Eds.), Intelligent Tutoring Systems - 4th International Conference, ITS 1998, Proceedings (pp. 384-393). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1452). Springer.

Webb, Geoffrey I. ; Kuzmycz, Mark. / Evaluation of data aging : A technique for discounting old data during student modeling. Intelligent Tutoring Systems - 4th International Conference, ITS 1998, Proceedings. editor / Carol L. Redfield ; Barry P. Goettl ; Valerie J. Shute ; Henry M. Halff. Springer, 1998. pp. 384-393 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{e39c572946a849cca033e7a45b00d34e,

title = "Evaluation of data aging: A technique for discounting old data during student modeling",

abstract = "Student modeling systems must operate in an environment in which a student{\textquoteright}s mastery of a subject matter is likely to change as a lesson progresses. A student model is formed from evaluation of evidence about the student{\textquoteright}s mastery of the domain. However, given that such mastery will change, older evidence is likely to be less valuable than recent evidence. Data aging addresses this issue by discounting the value of older evidence. This paper provides experimental evaluation of the effects of data aging. While it is demonstrated that data aging can result in statistically significant increases in both the number and accuracy of predictions that a modeling system makes, it is also demonstrated that the reverse can be true. Further, the effects experienced are of only small magnitude. It is argued that these results demonstrate some potential for data aging as a general strategy, but do not warrant employing data aging in its current form.",

author = "Webb, \{Geoffrey I.\} and Mark Kuzmycz",

year = "1998",

month = jan,

day = "1",

language = "English",

isbn = "3540647708",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer",

pages = "384--393",

editor = "Redfield, \{Carol L.\} and Goettl, \{Barry P.\} and Shute, \{Valerie J.\} and Halff, \{Henry M.\}",

booktitle = "Intelligent Tutoring Systems - 4th International Conference, ITS 1998, Proceedings",

address = "Switzerland",

note = "International Conference on Intelligent Tutoring Systems 1998, ITS 1998 ; Conference date: 16-08-1998 Through 19-08-1998",

}

Webb, GI & Kuzmycz, M 1998, Evaluation of data aging: A technique for discounting old data during student modeling. in CL Redfield, BP Goettl, VJ Shute & HM Halff (eds), Intelligent Tutoring Systems - 4th International Conference, ITS 1998, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 1452, Springer, pp. 384-393, International Conference on Intelligent Tutoring Systems 1998, San Antonio, United States of America, 16/08/98.

Evaluation of data aging: A technique for discounting old data during student modeling. / Webb, Geoffrey I.; Kuzmycz, Mark.
Intelligent Tutoring Systems - 4th International Conference, ITS 1998, Proceedings. ed. / Carol L. Redfield; Barry P. Goettl; Valerie J. Shute; Henry M. Halff. Springer, 1998. p. 384-393 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1452).

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › Research › peer-review

TY - GEN

T1 - Evaluation of data aging

T2 - International Conference on Intelligent Tutoring Systems 1998

AU - Webb, Geoffrey I.

AU - Kuzmycz, Mark

N1 - Conference code: 4th

PY - 1998/1/1

Y1 - 1998/1/1

N2 - Student modeling systems must operate in an environment in which a student’s mastery of a subject matter is likely to change as a lesson progresses. A student model is formed from evaluation of evidence about the student’s mastery of the domain. However, given that such mastery will change, older evidence is likely to be less valuable than recent evidence. Data aging addresses this issue by discounting the value of older evidence. This paper provides experimental evaluation of the effects of data aging. While it is demonstrated that data aging can result in statistically significant increases in both the number and accuracy of predictions that a modeling system makes, it is also demonstrated that the reverse can be true. Further, the effects experienced are of only small magnitude. It is argued that these results demonstrate some potential for data aging as a general strategy, but do not warrant employing data aging in its current form.

AB - Student modeling systems must operate in an environment in which a student’s mastery of a subject matter is likely to change as a lesson progresses. A student model is formed from evaluation of evidence about the student’s mastery of the domain. However, given that such mastery will change, older evidence is likely to be less valuable than recent evidence. Data aging addresses this issue by discounting the value of older evidence. This paper provides experimental evaluation of the effects of data aging. While it is demonstrated that data aging can result in statistically significant increases in both the number and accuracy of predictions that a modeling system makes, it is also demonstrated that the reverse can be true. Further, the effects experienced are of only small magnitude. It is argued that these results demonstrate some potential for data aging as a general strategy, but do not warrant employing data aging in its current form.

UR - https://www.scopus.com/pages/publications/84957064131

M3 - Conference Paper

AN - SCOPUS:84957064131

SN - 3540647708

SN - 9783540647706

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 384

EP - 393

BT - Intelligent Tutoring Systems - 4th International Conference, ITS 1998, Proceedings

A2 - Redfield, Carol L.

A2 - Goettl, Barry P.

A2 - Shute, Valerie J.

A2 - Halff, Henry M.

PB - Springer

Y2 - 16 August 1998 through 19 August 1998

ER -

Webb GI, Kuzmycz M. Evaluation of data aging: A technique for discounting old data during student modeling. In Redfield CL, Goettl BP, Shute VJ, Halff HM, editors, Intelligent Tutoring Systems - 4th International Conference, ITS 1998, Proceedings. Springer. 1998. p. 384-393. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

Evaluation of data aging: A technique for discounting old data during student modeling

Abstract

Publication series

Conference

Other files and links

Cite this