Polarity classification using structure-based vector representations of text

Alexander Hogenboom, Flavius Frasincar, Franciska M.G. de Jong, Uzay Kaymak

Research output: Contribution to journalArticleAcademicpeer-review

12 Citations (Scopus)

Abstract

The exploitation of structural aspects of content is becoming increasingly popular in rule-based polarity classification systems. Such systems typically weight the sentiment conveyed by text segments in accordance with these segments' roles in the structure of a text, as identified by deep linguistic processing. Conversely, state-of-the-art machine learning polarity classifiers typically aim to exploit patterns in vector representations of texts, mostly covering the occurrence of words or word groups in these texts. However, since structural aspects of content have been shown to contain valuable information as well, we propose to use structure-based features in vector representations of text. We evaluate the usefulness of our novel features on collections of English reviews in various domains. Our experimental results suggest that, even though word-based features are indispensable to good polarity classifiers, structure-based sentiment information provides valuable additional guidance that can help significantly improve the polarity classification performance of machine learning classifiers. The most informative features capture the sentiment conveyed by specific rhetorical elements that constitute a text's core or provide crucial contextual information.
Original languageUndefined
Pages (from-to)46-56
Number of pages11
JournalDecision support systems
Volume74
Issue numberJune
DOIs
Publication statusPublished - Jun 2015

Keywords

  • EWI-25971
  • METIS-312577
  • IR-96333

Cite this

Hogenboom, Alexander ; Frasincar, Flavius ; de Jong, Franciska M.G. ; Kaymak, Uzay. / Polarity classification using structure-based vector representations of text. In: Decision support systems. 2015 ; Vol. 74, No. June. pp. 46-56.
@article{07d4accbaebc4303a000fe492ace44d0,
title = "Polarity classification using structure-based vector representations of text",
abstract = "The exploitation of structural aspects of content is becoming increasingly popular in rule-based polarity classification systems. Such systems typically weight the sentiment conveyed by text segments in accordance with these segments' roles in the structure of a text, as identified by deep linguistic processing. Conversely, state-of-the-art machine learning polarity classifiers typically aim to exploit patterns in vector representations of texts, mostly covering the occurrence of words or word groups in these texts. However, since structural aspects of content have been shown to contain valuable information as well, we propose to use structure-based features in vector representations of text. We evaluate the usefulness of our novel features on collections of English reviews in various domains. Our experimental results suggest that, even though word-based features are indispensable to good polarity classifiers, structure-based sentiment information provides valuable additional guidance that can help significantly improve the polarity classification performance of machine learning classifiers. The most informative features capture the sentiment conveyed by specific rhetorical elements that constitute a text's core or provide crucial contextual information.",
keywords = "EWI-25971, METIS-312577, IR-96333",
author = "Alexander Hogenboom and Flavius Frasincar and {de Jong}, {Franciska M.G.} and Uzay Kaymak",
note = "eemcs-eprint-25971",
year = "2015",
month = "6",
doi = "10.1016/j.dss.2015.04.002",
language = "Undefined",
volume = "74",
pages = "46--56",
journal = "Decision support systems",
issn = "0167-9236",
publisher = "Elsevier",
number = "June",

}

Polarity classification using structure-based vector representations of text. / Hogenboom, Alexander; Frasincar, Flavius; de Jong, Franciska M.G.; Kaymak, Uzay.

In: Decision support systems, Vol. 74, No. June, 06.2015, p. 46-56.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - Polarity classification using structure-based vector representations of text

AU - Hogenboom, Alexander

AU - Frasincar, Flavius

AU - de Jong, Franciska M.G.

AU - Kaymak, Uzay

N1 - eemcs-eprint-25971

PY - 2015/6

Y1 - 2015/6

N2 - The exploitation of structural aspects of content is becoming increasingly popular in rule-based polarity classification systems. Such systems typically weight the sentiment conveyed by text segments in accordance with these segments' roles in the structure of a text, as identified by deep linguistic processing. Conversely, state-of-the-art machine learning polarity classifiers typically aim to exploit patterns in vector representations of texts, mostly covering the occurrence of words or word groups in these texts. However, since structural aspects of content have been shown to contain valuable information as well, we propose to use structure-based features in vector representations of text. We evaluate the usefulness of our novel features on collections of English reviews in various domains. Our experimental results suggest that, even though word-based features are indispensable to good polarity classifiers, structure-based sentiment information provides valuable additional guidance that can help significantly improve the polarity classification performance of machine learning classifiers. The most informative features capture the sentiment conveyed by specific rhetorical elements that constitute a text's core or provide crucial contextual information.

AB - The exploitation of structural aspects of content is becoming increasingly popular in rule-based polarity classification systems. Such systems typically weight the sentiment conveyed by text segments in accordance with these segments' roles in the structure of a text, as identified by deep linguistic processing. Conversely, state-of-the-art machine learning polarity classifiers typically aim to exploit patterns in vector representations of texts, mostly covering the occurrence of words or word groups in these texts. However, since structural aspects of content have been shown to contain valuable information as well, we propose to use structure-based features in vector representations of text. We evaluate the usefulness of our novel features on collections of English reviews in various domains. Our experimental results suggest that, even though word-based features are indispensable to good polarity classifiers, structure-based sentiment information provides valuable additional guidance that can help significantly improve the polarity classification performance of machine learning classifiers. The most informative features capture the sentiment conveyed by specific rhetorical elements that constitute a text's core or provide crucial contextual information.

KW - EWI-25971

KW - METIS-312577

KW - IR-96333

U2 - 10.1016/j.dss.2015.04.002

DO - 10.1016/j.dss.2015.04.002

M3 - Article

VL - 74

SP - 46

EP - 56

JO - Decision support systems

JF - Decision support systems

SN - 0167-9236

IS - June

ER -