The impact of positive, negative and topical relevance feedback

Rianne Kaptein, Jaap Kamps, Djoerd Hiemstra

Research output: Contribution to conferencePaperAcademic

55 Downloads (Pure)

Abstract

This document contains a description of experiments for the 2008 Relevance Feedback track. We experiment with different amounts of feedback, including negative relevance feedback. Feedback is implemented using massive weighted query expansion. Parsimonious query expansion using only relevant documents and Jelinek-Mercer smoothing performs best on this relevance feedback track dataset. Additional blind feedback gives better results, except when the blind feedback set is of the same size as the explicit feedback set. On a small number of topics topical feedback is applied, which turns out to be mainly beneficial for early precision.
Original languageUndefined
Number of pages7
Publication statusPublished - Nov 2008
EventSeventeenth Text REtrieval Conference, TREC-17 2008 - Gaithersburg, United States
Duration: 18 Nov 200821 Nov 2008
Conference number: 17

Conference

ConferenceSeventeenth Text REtrieval Conference, TREC-17 2008
Abbreviated titleTREC
CountryUnited States
CityGaithersburg
Period18/11/0821/11/08

Keywords

  • EWI-23223
  • IR-85286

Cite this

Kaptein, R., Kamps, J., & Hiemstra, D. (2008). The impact of positive, negative and topical relevance feedback. Paper presented at Seventeenth Text REtrieval Conference, TREC-17 2008, Gaithersburg, United States.
Kaptein, Rianne ; Kamps, Jaap ; Hiemstra, Djoerd. / The impact of positive, negative and topical relevance feedback. Paper presented at Seventeenth Text REtrieval Conference, TREC-17 2008, Gaithersburg, United States.7 p.
@conference{0267dbb63f214af6b89a9fdf8ef7730c,
title = "The impact of positive, negative and topical relevance feedback",
abstract = "This document contains a description of experiments for the 2008 Relevance Feedback track. We experiment with different amounts of feedback, including negative relevance feedback. Feedback is implemented using massive weighted query expansion. Parsimonious query expansion using only relevant documents and Jelinek-Mercer smoothing performs best on this relevance feedback track dataset. Additional blind feedback gives better results, except when the blind feedback set is of the same size as the explicit feedback set. On a small number of topics topical feedback is applied, which turns out to be mainly beneficial for early precision.",
keywords = "EWI-23223, IR-85286",
author = "Rianne Kaptein and Jaap Kamps and Djoerd Hiemstra",
year = "2008",
month = "11",
language = "Undefined",
note = "null ; Conference date: 18-11-2008 Through 21-11-2008",

}

Kaptein, R, Kamps, J & Hiemstra, D 2008, 'The impact of positive, negative and topical relevance feedback' Paper presented at Seventeenth Text REtrieval Conference, TREC-17 2008, Gaithersburg, United States, 18/11/08 - 21/11/08, .

The impact of positive, negative and topical relevance feedback. / Kaptein, Rianne; Kamps, Jaap; Hiemstra, Djoerd.

2008. Paper presented at Seventeenth Text REtrieval Conference, TREC-17 2008, Gaithersburg, United States.

Research output: Contribution to conferencePaperAcademic

TY - CONF

T1 - The impact of positive, negative and topical relevance feedback

AU - Kaptein, Rianne

AU - Kamps, Jaap

AU - Hiemstra, Djoerd

PY - 2008/11

Y1 - 2008/11

N2 - This document contains a description of experiments for the 2008 Relevance Feedback track. We experiment with different amounts of feedback, including negative relevance feedback. Feedback is implemented using massive weighted query expansion. Parsimonious query expansion using only relevant documents and Jelinek-Mercer smoothing performs best on this relevance feedback track dataset. Additional blind feedback gives better results, except when the blind feedback set is of the same size as the explicit feedback set. On a small number of topics topical feedback is applied, which turns out to be mainly beneficial for early precision.

AB - This document contains a description of experiments for the 2008 Relevance Feedback track. We experiment with different amounts of feedback, including negative relevance feedback. Feedback is implemented using massive weighted query expansion. Parsimonious query expansion using only relevant documents and Jelinek-Mercer smoothing performs best on this relevance feedback track dataset. Additional blind feedback gives better results, except when the blind feedback set is of the same size as the explicit feedback set. On a small number of topics topical feedback is applied, which turns out to be mainly beneficial for early precision.

KW - EWI-23223

KW - IR-85286

M3 - Paper

ER -

Kaptein R, Kamps J, Hiemstra D. The impact of positive, negative and topical relevance feedback. 2008. Paper presented at Seventeenth Text REtrieval Conference, TREC-17 2008, Gaithersburg, United States.