Score Normalization using Logistic Regression with Expected Parameters

Robin Aly

Abstract

State-of-the-art score normalization methods use generative models that rely on sometimes unrealistic assumptions. We propose a novel parameter estimation method for score normalization based on logistic regression. Experiments on the Gov2 and CluewebA collection indicate that our method is consistently more precise in predicting the number of relevant documents in the top-n ranks compared to a state-of- the-art generative approach and another parameter estimate for logistic regression.
Original languageUndefined
Title of host publicationProceedings of the 36th European Conference on IR Research, ECIR 2014
Place of PublicationLondon
PublisherSpringer Verlag
Pages579-584
Number of pages6
ISBN (Print)978-3-319-06027-9
DOIs
StatePublished - Apr 2014
Event36th European Conference on Information Retrieval, ECIR 2014 - Amsterdam, Netherlands

Publication series

NameLecture Notes in Computer Science
PublisherSpringer Verlag
Volume8416
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference36th European Conference on Information Retrieval, ECIR 2014
Abbreviated titleECIR
CountryNetherlands
CityAmsterdam
Period13/04/1416/04/14

Fingerprint

Generative
Normalization
Logistic regression
State of the art
Experiment
Parameter estimation

Keywords

  • METIS-312536
  • EWI-25891
  • IR-95306

Cite this

Aly, R. (2014). Score Normalization using Logistic Regression with Expected Parameters. In Proceedings of the 36th European Conference on IR Research, ECIR 2014 (pp. 579-584). (Lecture Notes in Computer Science; Vol. 8416). London: Springer Verlag. DOI: 10.1007/978-3-319-06028-6_60

Aly, Robin / Score Normalization using Logistic Regression with Expected Parameters.

Proceedings of the 36th European Conference on IR Research, ECIR 2014. London : Springer Verlag, 2014. p. 579-584 (Lecture Notes in Computer Science; Vol. 8416).

Research output: Scientific - peer-reviewConference contribution

@inbook{cbb507d8b9834e6b884080755a181971,
title = "Score Normalization using Logistic Regression with Expected Parameters",
abstract = "State-of-the-art score normalization methods use generative models that rely on sometimes unrealistic assumptions. We propose a novel parameter estimation method for score normalization based on logistic regression. Experiments on the Gov2 and CluewebA collection indicate that our method is consistently more precise in predicting the number of relevant documents in the top-n ranks compared to a state-of- the-art generative approach and another parameter estimate for logistic regression.",
keywords = "METIS-312536, EWI-25891, IR-95306",
author = "Robin Aly",
note = "10.1007/978-3-319-06028-6_60",
year = "2014",
month = "4",
doi = "10.1007/978-3-319-06028-6_60",
isbn = "978-3-319-06027-9",
series = "Lecture Notes in Computer Science",
publisher = "Springer Verlag",
pages = "579--584",
booktitle = "Proceedings of the 36th European Conference on IR Research, ECIR 2014",
address = "Germany",

}

Aly, R 2014, Score Normalization using Logistic Regression with Expected Parameters. in Proceedings of the 36th European Conference on IR Research, ECIR 2014. Lecture Notes in Computer Science, vol. 8416, Springer Verlag, London, pp. 579-584, 36th European Conference on Information Retrieval, ECIR 2014, Amsterdam, Netherlands, 13-16 April. DOI: 10.1007/978-3-319-06028-6_60

Score Normalization using Logistic Regression with Expected Parameters. / Aly, Robin.

Proceedings of the 36th European Conference on IR Research, ECIR 2014. London : Springer Verlag, 2014. p. 579-584 (Lecture Notes in Computer Science; Vol. 8416).

Research output: Scientific - peer-reviewConference contribution

TY - CHAP

T1 - Score Normalization using Logistic Regression with Expected Parameters

AU - Aly,Robin

N1 - 10.1007/978-3-319-06028-6_60

PY - 2014/4

Y1 - 2014/4

N2 - State-of-the-art score normalization methods use generative models that rely on sometimes unrealistic assumptions. We propose a novel parameter estimation method for score normalization based on logistic regression. Experiments on the Gov2 and CluewebA collection indicate that our method is consistently more precise in predicting the number of relevant documents in the top-n ranks compared to a state-of- the-art generative approach and another parameter estimate for logistic regression.

AB - State-of-the-art score normalization methods use generative models that rely on sometimes unrealistic assumptions. We propose a novel parameter estimation method for score normalization based on logistic regression. Experiments on the Gov2 and CluewebA collection indicate that our method is consistently more precise in predicting the number of relevant documents in the top-n ranks compared to a state-of- the-art generative approach and another parameter estimate for logistic regression.

KW - METIS-312536

KW - EWI-25891

KW - IR-95306

U2 - 10.1007/978-3-319-06028-6_60

DO - 10.1007/978-3-319-06028-6_60

M3 - Conference contribution

SN - 978-3-319-06027-9

T3 - Lecture Notes in Computer Science

SP - 579

EP - 584

BT - Proceedings of the 36th European Conference on IR Research, ECIR 2014

PB - Springer Verlag

ER -

Aly R. Score Normalization using Logistic Regression with Expected Parameters. In Proceedings of the 36th European Conference on IR Research, ECIR 2014. London: Springer Verlag. 2014. p. 579-584. (Lecture Notes in Computer Science). Available from, DOI: 10.1007/978-3-319-06028-6_60