Training data cleaning for text classification

Andrea Esuli, Fabrizio Sebastiani

Research output: Chapter in Book/Report/Conference proceedingConference contribution

11 Citations (Scopus)

Abstract

In text classification (TC) and other tasks involving supervised learning, labelled data may be scarce or expensive to obtain; strategies are thus needed for maximizing the effectiveness of the resulting classifiers while minimizing the required amount of training effort. Training data cleaning (TDC) consists in devising ranking functions that sort the original training examples in terms of how likely it is that the human annotator has misclassified them, thereby providing a convenient means for the human annotator to revise the training set so as to improve its quality. Working in the context of boosting-based learning methods we present three different techniques for performing TDC and, on two widely used TC benchmarks, evaluate them by their capability of spotting misclassified texts purposefully inserted in the training set.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages29-41
Number of pages13
Volume5766 LNCS
DOIs
Publication statusPublished - 2009
Externally publishedYes
Event2nd International Conference on the Theory of Information Retrieval, ICTIR 2009 - Cambridge
Duration: 10 Sep 200912 Sep 2009

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume5766 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other2nd International Conference on the Theory of Information Retrieval, ICTIR 2009
CityCambridge
Period10/9/0912/9/09

Fingerprint

Text Classification
Cleaning
Supervised learning
Classifiers
Ranking Function
Supervised Learning
Boosting
Sort
Training
Likely
Classifier
Benchmark
Evaluate

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Esuli, A., & Sebastiani, F. (2009). Training data cleaning for text classification. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 5766 LNCS, pp. 29-41). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5766 LNCS). https://doi.org/10.1007/978-3-642-04417-5_4

Training data cleaning for text classification. / Esuli, Andrea; Sebastiani, Fabrizio.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 5766 LNCS 2009. p. 29-41 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5766 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Esuli, A & Sebastiani, F 2009, Training data cleaning for text classification. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 5766 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 5766 LNCS, pp. 29-41, 2nd International Conference on the Theory of Information Retrieval, ICTIR 2009, Cambridge, 10/9/09. https://doi.org/10.1007/978-3-642-04417-5_4
Esuli A, Sebastiani F. Training data cleaning for text classification. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 5766 LNCS. 2009. p. 29-41. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-642-04417-5_4
Esuli, Andrea ; Sebastiani, Fabrizio. / Training data cleaning for text classification. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 5766 LNCS 2009. pp. 29-41 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{e558b0e2485b43029fcae4d2bd0c49bf,
title = "Training data cleaning for text classification",
abstract = "In text classification (TC) and other tasks involving supervised learning, labelled data may be scarce or expensive to obtain; strategies are thus needed for maximizing the effectiveness of the resulting classifiers while minimizing the required amount of training effort. Training data cleaning (TDC) consists in devising ranking functions that sort the original training examples in terms of how likely it is that the human annotator has misclassified them, thereby providing a convenient means for the human annotator to revise the training set so as to improve its quality. Working in the context of boosting-based learning methods we present three different techniques for performing TDC and, on two widely used TC benchmarks, evaluate them by their capability of spotting misclassified texts purposefully inserted in the training set.",
author = "Andrea Esuli and Fabrizio Sebastiani",
year = "2009",
doi = "10.1007/978-3-642-04417-5_4",
language = "English",
isbn = "3642044166",
volume = "5766 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "29--41",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Training data cleaning for text classification

AU - Esuli, Andrea

AU - Sebastiani, Fabrizio

PY - 2009

Y1 - 2009

N2 - In text classification (TC) and other tasks involving supervised learning, labelled data may be scarce or expensive to obtain; strategies are thus needed for maximizing the effectiveness of the resulting classifiers while minimizing the required amount of training effort. Training data cleaning (TDC) consists in devising ranking functions that sort the original training examples in terms of how likely it is that the human annotator has misclassified them, thereby providing a convenient means for the human annotator to revise the training set so as to improve its quality. Working in the context of boosting-based learning methods we present three different techniques for performing TDC and, on two widely used TC benchmarks, evaluate them by their capability of spotting misclassified texts purposefully inserted in the training set.

AB - In text classification (TC) and other tasks involving supervised learning, labelled data may be scarce or expensive to obtain; strategies are thus needed for maximizing the effectiveness of the resulting classifiers while minimizing the required amount of training effort. Training data cleaning (TDC) consists in devising ranking functions that sort the original training examples in terms of how likely it is that the human annotator has misclassified them, thereby providing a convenient means for the human annotator to revise the training set so as to improve its quality. Working in the context of boosting-based learning methods we present three different techniques for performing TDC and, on two widely used TC benchmarks, evaluate them by their capability of spotting misclassified texts purposefully inserted in the training set.

UR - http://www.scopus.com/inward/record.url?scp=70350614398&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=70350614398&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-04417-5_4

DO - 10.1007/978-3-642-04417-5_4

M3 - Conference contribution

SN - 3642044166

SN - 9783642044168

VL - 5766 LNCS

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

SP - 29

EP - 41

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

ER -