AWESSOME: An unsupervised sentiment intensity scoring framework using neural word embeddings

Amal Htait; Leif Azzopardi

doi:10.1007/978-3-030-72240-1_56

AWESSOME: An unsupervised sentiment intensity scoring framework using neural word embeddings

Amal Htait, Leif Azzopardi

Computer And Information Sciences

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution book

7 Downloads (Pure)

Abstract

Sentiment analysis (SA) is the key element for a variety of opinion and attitude mining tasks. While various unsupervised SA tools already exist, a central problem is that they are lexicon-based where the lexicons used are limited, leading to a vocabulary mismatch. In this paper, we present an unsupervised word embedding-based sentiment scoring framework for sentiment intensity scoring (SIS). The framework generalizes and combines past works so that pre-existing lexicons (e.g. VADER, LabMT) and word embeddings (e.g. BERT, RoBERTa) can be used to address this problem, with no require training, and while providing fine grained SIS of words and phrases. The framework is scalable and extensible, so that custom lexicons or word embeddings can be used to core methods, and to even create new corpus specific lexicons without the need for extensive supervised learning and retraining. The Python 3 toolkit is open source, freely available from GitHub and can be directly installed via
pip install awessome.

Original language	English
Title of host publication	Advances in Information Retrieval. ECIR 2021. Lecture Notes in Computer Science
Editors	Djoerd Hiemstra, Marie-Francine Moens, Josiane Mothe, Raffaele Perego, Martin Potthast, Fabrizio Sebastiani
Place of Publication	Cham, Switzerland
Publisher	Springer
Chapter	56
Pages	509-513
Number of pages	5
Volume	12657
ISBN (Electronic)	9783030722401
ISBN (Print)	9783030722395
DOIs	https://doi.org/10.1007/978-3-030-72240-1_56
Publication status	Published - 28 Mar 2021
Event	European Conference on Information Retrieval 2021 - Online, Lucca, Italy Duration: 28 Mar 2021 → 1 Apr 2021 Conference number: 43 https://www.ecir2021.eu/

Publication series

Name	Lecture Notes in Computer Science (LNCS)
Publisher	Springer International
ISSN (Print)	0302-9743

Conference

Conference	European Conference on Information Retrieval 2021
Abbreviated title	ECIR 2021
Country	Italy
City	Lucca
Period	28/03/21 → 1/04/21
Internet address	https://www.ecir2021.eu/

Keywords

sentiment intensity
pre-trained language model
lexicon
BERT
VADER

Access to Document

10.1007/978-3-030-72240-1_56

Htait-Azzopardi-ECIR-2021-AWESSOME-an-unsupervised-sentiment-intensity-scoring-frameworkAccepted author manuscript, 215 KB

Fingerprint Dive into the research topics of 'AWESSOME: An unsupervised sentiment intensity scoring framework using neural word embeddings'. Together they form a unique fingerprint.

1 Active

Cumulative Revelations in Personal Data
Azzopardi, L.
EPSRC (Engineering and Physical Sciences Research Council)
1/04/19 → 31/03/22
Project: Research

Cite this

Htait, A., & Azzopardi, L. (2021). AWESSOME: An unsupervised sentiment intensity scoring framework using neural word embeddings. In D. Hiemstra, M-F. Moens, J. Mothe, R. Perego, M. Potthast, & F. Sebastiani (Eds.), Advances in Information Retrieval. ECIR 2021. Lecture Notes in Computer Science (Vol. 12657, pp. 509-513). (Lecture Notes in Computer Science (LNCS)). Springer. https://doi.org/10.1007/978-3-030-72240-1_56

Htait, Amal ; Azzopardi, Leif. / AWESSOME : An unsupervised sentiment intensity scoring framework using neural word embeddings. Advances in Information Retrieval. ECIR 2021. Lecture Notes in Computer Science. editor / Djoerd Hiemstra ; Marie-Francine Moens ; Josiane Mothe ; Raffaele Perego ; Martin Potthast ; Fabrizio Sebastiani. Vol. 12657 Cham, Switzerland : Springer, 2021. pp. 509-513 (Lecture Notes in Computer Science (LNCS)).

@inproceedings{6cbf1ffd446847bca4e0d881fd41d168,

title = "AWESSOME: An unsupervised sentiment intensity scoring framework using neural word embeddings",

abstract = "Sentiment analysis (SA) is the key element for a variety of opinion and attitude mining tasks. While various unsupervised SA tools already exist, a central problem is that they are lexicon-based where the lexicons used are limited, leading to a vocabulary mismatch. In this paper, we present an unsupervised word embedding-based sentiment scoring framework for sentiment intensity scoring (SIS). The framework generalizes and combines past works so that pre-existing lexicons (e.g. VADER, LabMT) and word embeddings (e.g. BERT, RoBERTa) can be used to address this problem, with no require training, and while providing fine grained SIS of words and phrases. The framework is scalable and extensible, so that custom lexicons or word embeddings can be used to core methods, and to even create new corpus specific lexicons without the need for extensive supervised learning and retraining. The Python 3 toolkit is open source, freely available from GitHub and can be directly installed via pip install awessome.",

keywords = "sentiment intensity, pre-trained language model, lexicon, BERT, VADER",

author = "Amal Htait and Leif Azzopardi",

year = "2021",

month = mar,

day = "28",

doi = "10.1007/978-3-030-72240-1_56",

language = "English",

isbn = "9783030722395",

volume = "12657",

series = "Lecture Notes in Computer Science (LNCS)",

publisher = "Springer",

pages = "509--513",

editor = "Djoerd Hiemstra and Marie-Francine Moens and Josiane Mothe and Raffaele Perego and Martin Potthast and Fabrizio Sebastiani",

booktitle = "Advances in Information Retrieval. ECIR 2021. Lecture Notes in Computer Science",

note = "European Conference on Information Retrieval 2021, ECIR 2021 ; Conference date: 28-03-2021 Through 01-04-2021",

url = "https://www.ecir2021.eu/",

}

Htait, A & Azzopardi, L 2021, AWESSOME: An unsupervised sentiment intensity scoring framework using neural word embeddings. in D Hiemstra, M-F Moens, J Mothe, R Perego, M Potthast & F Sebastiani (eds), Advances in Information Retrieval. ECIR 2021. Lecture Notes in Computer Science. vol. 12657, Lecture Notes in Computer Science (LNCS), Springer, Cham, Switzerland, pp. 509-513, European Conference on Information Retrieval 2021, Lucca, Italy, 28/03/21. https://doi.org/10.1007/978-3-030-72240-1_56

AWESSOME : An unsupervised sentiment intensity scoring framework using neural word embeddings. / Htait, Amal ; Azzopardi, Leif.

Advances in Information Retrieval. ECIR 2021. Lecture Notes in Computer Science. ed. / Djoerd Hiemstra; Marie-Francine Moens; Josiane Mothe; Raffaele Perego; Martin Potthast; Fabrizio Sebastiani. Vol. 12657 Cham, Switzerland : Springer, 2021. p. 509-513 (Lecture Notes in Computer Science (LNCS)).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution book

TY - GEN

T1 - AWESSOME

T2 - European Conference on Information Retrieval 2021

AU - Htait, Amal

AU - Azzopardi, Leif

N1 - Conference code: 43

PY - 2021/3/28

Y1 - 2021/3/28

N2 - Sentiment analysis (SA) is the key element for a variety of opinion and attitude mining tasks. While various unsupervised SA tools already exist, a central problem is that they are lexicon-based where the lexicons used are limited, leading to a vocabulary mismatch. In this paper, we present an unsupervised word embedding-based sentiment scoring framework for sentiment intensity scoring (SIS). The framework generalizes and combines past works so that pre-existing lexicons (e.g. VADER, LabMT) and word embeddings (e.g. BERT, RoBERTa) can be used to address this problem, with no require training, and while providing fine grained SIS of words and phrases. The framework is scalable and extensible, so that custom lexicons or word embeddings can be used to core methods, and to even create new corpus specific lexicons without the need for extensive supervised learning and retraining. The Python 3 toolkit is open source, freely available from GitHub and can be directly installed via pip install awessome.

AB - Sentiment analysis (SA) is the key element for a variety of opinion and attitude mining tasks. While various unsupervised SA tools already exist, a central problem is that they are lexicon-based where the lexicons used are limited, leading to a vocabulary mismatch. In this paper, we present an unsupervised word embedding-based sentiment scoring framework for sentiment intensity scoring (SIS). The framework generalizes and combines past works so that pre-existing lexicons (e.g. VADER, LabMT) and word embeddings (e.g. BERT, RoBERTa) can be used to address this problem, with no require training, and while providing fine grained SIS of words and phrases. The framework is scalable and extensible, so that custom lexicons or word embeddings can be used to core methods, and to even create new corpus specific lexicons without the need for extensive supervised learning and retraining. The Python 3 toolkit is open source, freely available from GitHub and can be directly installed via pip install awessome.

KW - sentiment intensity

KW - pre-trained language model

KW - lexicon

KW - BERT

KW - VADER

U2 - 10.1007/978-3-030-72240-1_56

DO - 10.1007/978-3-030-72240-1_56

M3 - Conference contribution book

SN - 9783030722395

VL - 12657

T3 - Lecture Notes in Computer Science (LNCS)

SP - 509

EP - 513

BT - Advances in Information Retrieval. ECIR 2021. Lecture Notes in Computer Science

A2 - Hiemstra, Djoerd

A2 - Moens, Marie-Francine

A2 - Mothe, Josiane

A2 - Perego, Raffaele

A2 - Potthast, Martin

A2 - Sebastiani, Fabrizio

PB - Springer

CY - Cham, Switzerland

Y2 - 28 March 2021 through 1 April 2021

ER -

Htait A , Azzopardi L. AWESSOME: An unsupervised sentiment intensity scoring framework using neural word embeddings. In Hiemstra D, Moens M-F, Mothe J, Perego R, Potthast M, Sebastiani F, editors, Advances in Information Retrieval. ECIR 2021. Lecture Notes in Computer Science. Vol. 12657. Cham, Switzerland: Springer. 2021. p. 509-513. (Lecture Notes in Computer Science (LNCS)). https://doi.org/10.1007/978-3-030-72240-1_56

AWESSOME: An unsupervised sentiment intensity scoring framework using neural word embeddings

Abstract

Publication series

Conference

Keywords

Access to Document

Cumulative Revelations in Personal Data

Cite this