Improving the affective analysis in texts. Automatic method to detect affective intensity in lexicons based on Plutchik’s wheel of emotions
| Published date | 09 December 2019 |
| Pages | 984-1006 |
| DOI | https://doi.org/10.1108/EL-11-2018-0219 |
| Date | 09 December 2019 |
| Author | Carlos Molina Beltrán,Alejandra Andrea Segura Navarrete,Christian Vidal-Castro,Clemente Rubio-Manzano,Claudia Martínez-Araneda |
| Subject Matter | Information & knowledge management |
Improving the affective analysis
in texts
Automatic method to detect affective intensity
in lexicons based on Plutchik’s wheel
of emotions
Carlos Molina Beltrán,Alejandra Andrea Segura Navarrete and
Christian Vidal-Castro
Departamento de Sistemas de Informaci
on, Universidad del Bío-Bío,
Concepci
on, Chile
Clemente Rubio-Manzano
Departamento de Sistemas de Informaci
on, Universidad del Bío-Bío, Concepci
on,
Chile and Departamento de Matemáticas, Universidad de Cádiz, Cádiz, Spain, and
Claudia Martínez-Araneda
Department of Computer Science, Universidad Cat
olica de la SantísimaConcepci
on,
Concepci
on, Chile
Abstract
Purpose –This paper aims to propose a method for automatically labelling an affective lexicon with
intensity values by using the WordNet Similarity(WS) software package with the purpose of improving the
results of an affective analysis process, which is relevant to interpreting the textual information that is
available in socialnetworks. The hypothesis states that it is possible to improveaffective analysis by using a
lexicon that is enriched with the intensityvalues obtained from similarity metrics. Encouraging results were
obtained when an affective analysis basedon a labelled lexicon was compared with that based on another
lexiconwithout intensity values.
Design/methodology/approach –The authors propose a method for the automatic extraction of the
affective intensityvalues of words using the similarity metrics implemented in WS. First, the intensityvalues
were calculated for words having an affective root in WordNet. Then, to evaluate the effectiveness of the
proposal, the results of the affectiveanalysis based on a labelled lexicon were compared to the results of an
analysiswith and without affective intensity values.
Findings –The main contribution of this researchis a method for the automatic extraction of the intensity
values of affective words used to enrich a lexicon compared with the manual labelling process. The results
obtained from the affective analysis with the new lexicon are encouraging, as they provide a better
performancethan those achieved using a lexicon withoutaffective intensity values.
Research limitations/implications –Given the restrictionsfor calculating the similarity between two
words, the lexicon labelledwith intensity values is a subset of the original lexicon,which means that a large
proportionof the words in the corpus are not labelled in the new lexicon.
Practical implications –The practical implicationsof this work include providing tools to improve the
analysis of the feelings of the users of socialnetworks. In particular, it is of interest to provide an affective
This paper is the result of work by the SOMOS research group (SOftware - MOdelling - Science),
funded by the Direcci
on de Investigaci
on and Facultad de Ciencias Empresariales of the Universidad
del Bío-Bío, Chile. The authors thank the Facultad de Ingeniería de la Universidad Cat
olica de la
Santísima Concepci
on, Chile.
EL
37,6
984
Received9 November 2018
Revised14 March 2019
14May 2019
29July 2019
Accepted2 September 2019
TheElectronic Library
Vol.37 No. 6, 2019
pp. 984-1006
© Emerald Publishing Limited
0264-0473
DOI 10.1108/EL-11-2018-0219
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/0264-0473.htm
lexicon that improves attempts to solve the problems of a digital society, such as the detection of
cyberbullying.In this case, by achieving greaterprecision in the detection of emotions, it is possible to detect
the roles of participants in a situationof cyberbullying, for example, the bully and victim. Other problems in
which the applicationof affective lexicons is of importance are the detectionof aggressiveness against women
or gender violenceor the detection of depressive states in young people and children.
Social implications –This work is interestedin providing an affective lexicon that improvesattempts to
solve the problems of a digital society, such as the detection of cyberbullying. In this case, by achieving
greater precision in the detectionof emotions, it is possible to detect the roles of participants in a situation of
cyber bullying, for example, the bully and victim. Other problems in which the application of affective
lexicons is of importance are the detection of aggressiveness against women or gender violence or the
detectionof depressive states in young people and children.
Originality/value –The originality of the researchlies in the proposed method for automatically labelling
the words of an affectivelexicon with intensity values by using WS. Todate, a lexicon labelled with intensity
values has been constructed using the opinionsof experts, but that method is more expensive and requires
more time than other existing methods.On the other hand, the new method developed herein is applicable to
larger lexicons,requires less time and facilitates automaticupdating.
Keywords Opinion mining, Affective analysis, Affective Lexicon, Similarity metrics
Paper type Research paper
Introduction
The subjectivity analysis of texts includes other, more specific analyses, such as opinion
mining, irony/sarcasm managementand affective analysis. Affective analysis includes a set
of techniques used to identify the emotions expressed in texts beyond positive, negative or
neutral polarity (Hull and Grefenstette, 1996;Shanahan et al., 2004). For example, in
marketing, to determine the level of acceptance of a brand or product (Trivedi and Dey,
2018), opinion mining canbe used to determine whether the comments related to the product
are positive, negative or neutral (Zhang et al.,2018), whereas affective analysis determines
the emotions contained in texts as expressed by consumers when they talk about a brand,
for example, assessing whether they are dominant emotions, such as anger, joy or disgust.
While both analysesare useful, affective analysis allows a much higher levelof detail, which
motivates its application in the detection of cyberbullying, depressive behaviour, gender
aggressiveness and suicidal tendencies, among other applications. Evidence collected from
psychological studiesshows that victims of cyberbullying experience a variety of emotional
stresses, highlighting an increase in emotional distress, anger and sadness (Ybarra and
Mitchell, 2004).
Since 2017, research on affective analysis has tried to increase the accuracy of the
analysis, based on the facts that texts can express more than one emotion at the same
time and that multiple words that evoke the same emotion may have different affective
intensity values. In this research, it is proposed that it is possible to improve the results
of an affective analysis based on a lexicon by using an enriched lexicon with the
affective intensity values of the words. A method is proposed to automatically extract
the affective intensity values of the words by using various similarity metrics (Figure 1).
The affective intensity value of a word is obtained from the similarity of this word to a
reference word (rw), and the rw is an affective word used as a pivot. Finally, to evaluate
the effectiveness of this proposal, the results of an affective analysis based on a lexicon
are compared to the results of analyses with and without intensity values using the
corpus published by Mohammad (2018).
The remainder of this article is organized as follows: the next section presents the
background and the related work on sentiment analysiswith an emphasis on lexicon-based
approaches. Section 3 describes the methodologyapplied, including a detailed description of
Affective
intensity in
lexicons
985
To continue reading
Request your trial