Feature selection for ordinal text classification

Stefano Baccianella, Andrea Esuli, Fabrizio Sebastiani

Research output: Contribution to journalArticle

30 Citations (Scopus)

Abstract

Ordinal classification (also known as ordinal regression) is a supervised learning task that consists of estimating the rating of a data item on a fixed, discrete rating scale. This problem is receiving increased attention from the sentiment analysis and opinion mining community due to the importance of automatically rating large amounts of product review data in digitalform. As in other supervised learning tasks such as binary or multiclass classification, feature selection is often needed in order to improve efficiency and avoid overfitting. However, although feature selection has been extensively studied for other classification tasks, it has not for ordinal classification. In this letter, we present six novel feature selectionmethods that we have specifically devised for ordinal classification and test them on two data sets of product review data against three methods previously known from the literature, using two learning algorithms from the support vector regression tradition. The experimental results show that all six proposed metrics largely outperform all three baseline techniques (and are more stable than these others by an order of magnitude), on both data sets and for both learning algorithms.

Original languageEnglish
Pages (from-to)557-591
Number of pages35
JournalNeural Computation
Volume26
Issue number3
DOIs
Publication statusPublished - 2014
Externally publishedYes

Fingerprint

Learning
Feature Selection
Efficiency
Datasets
Rating
Rating Scales
Sentiment

ASJC Scopus subject areas

  • Cognitive Neuroscience
  • Arts and Humanities (miscellaneous)

Cite this

Baccianella, S., Esuli, A., & Sebastiani, F. (2014). Feature selection for ordinal text classification. Neural Computation, 26(3), 557-591. https://doi.org/10.1162/NECO_a_00558

Feature selection for ordinal text classification. / Baccianella, Stefano; Esuli, Andrea; Sebastiani, Fabrizio.

In: Neural Computation, Vol. 26, No. 3, 2014, p. 557-591.

Research output: Contribution to journalArticle

Baccianella, S, Esuli, A & Sebastiani, F 2014, 'Feature selection for ordinal text classification', Neural Computation, vol. 26, no. 3, pp. 557-591. https://doi.org/10.1162/NECO_a_00558
Baccianella, Stefano ; Esuli, Andrea ; Sebastiani, Fabrizio. / Feature selection for ordinal text classification. In: Neural Computation. 2014 ; Vol. 26, No. 3. pp. 557-591.
@article{33fb7a371e7745d5bf1084f18870f0e0,
title = "Feature selection for ordinal text classification",
abstract = "Ordinal classification (also known as ordinal regression) is a supervised learning task that consists of estimating the rating of a data item on a fixed, discrete rating scale. This problem is receiving increased attention from the sentiment analysis and opinion mining community due to the importance of automatically rating large amounts of product review data in digitalform. As in other supervised learning tasks such as binary or multiclass classification, feature selection is often needed in order to improve efficiency and avoid overfitting. However, although feature selection has been extensively studied for other classification tasks, it has not for ordinal classification. In this letter, we present six novel feature selectionmethods that we have specifically devised for ordinal classification and test them on two data sets of product review data against three methods previously known from the literature, using two learning algorithms from the support vector regression tradition. The experimental results show that all six proposed metrics largely outperform all three baseline techniques (and are more stable than these others by an order of magnitude), on both data sets and for both learning algorithms.",
author = "Stefano Baccianella and Andrea Esuli and Fabrizio Sebastiani",
year = "2014",
doi = "10.1162/NECO_a_00558",
language = "English",
volume = "26",
pages = "557--591",
journal = "Neural Computation",
issn = "0899-7667",
publisher = "MIT Press Journals",
number = "3",

}

TY - JOUR

T1 - Feature selection for ordinal text classification

AU - Baccianella, Stefano

AU - Esuli, Andrea

AU - Sebastiani, Fabrizio

PY - 2014

Y1 - 2014

N2 - Ordinal classification (also known as ordinal regression) is a supervised learning task that consists of estimating the rating of a data item on a fixed, discrete rating scale. This problem is receiving increased attention from the sentiment analysis and opinion mining community due to the importance of automatically rating large amounts of product review data in digitalform. As in other supervised learning tasks such as binary or multiclass classification, feature selection is often needed in order to improve efficiency and avoid overfitting. However, although feature selection has been extensively studied for other classification tasks, it has not for ordinal classification. In this letter, we present six novel feature selectionmethods that we have specifically devised for ordinal classification and test them on two data sets of product review data against three methods previously known from the literature, using two learning algorithms from the support vector regression tradition. The experimental results show that all six proposed metrics largely outperform all three baseline techniques (and are more stable than these others by an order of magnitude), on both data sets and for both learning algorithms.

AB - Ordinal classification (also known as ordinal regression) is a supervised learning task that consists of estimating the rating of a data item on a fixed, discrete rating scale. This problem is receiving increased attention from the sentiment analysis and opinion mining community due to the importance of automatically rating large amounts of product review data in digitalform. As in other supervised learning tasks such as binary or multiclass classification, feature selection is often needed in order to improve efficiency and avoid overfitting. However, although feature selection has been extensively studied for other classification tasks, it has not for ordinal classification. In this letter, we present six novel feature selectionmethods that we have specifically devised for ordinal classification and test them on two data sets of product review data against three methods previously known from the literature, using two learning algorithms from the support vector regression tradition. The experimental results show that all six proposed metrics largely outperform all three baseline techniques (and are more stable than these others by an order of magnitude), on both data sets and for both learning algorithms.

UR - http://www.scopus.com/inward/record.url?scp=84893438758&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84893438758&partnerID=8YFLogxK

U2 - 10.1162/NECO_a_00558

DO - 10.1162/NECO_a_00558

M3 - Article

AN - SCOPUS:84893438758

VL - 26

SP - 557

EP - 591

JO - Neural Computation

JF - Neural Computation

SN - 0899-7667

IS - 3

ER -