helpfn

Received June 26, 2020, accepted July 12, 2020, date of publication July 15, 2020, date of current version July 31, 2020.

Digital Object Identifier 10.1109/ACCESS.2020.3009482

Opinion Mining, Sentiment Analysis and Emotion Understanding in Advertising: A Bibliometric Analysis PABLO SÁNCHEZ-NÚÑEZ 1, MANUEL J. COBO 2, CARLOS DE LAS HERAS-PEDROSA 3, JOSÉ IGNACIO PELÁEZ 4, AND ENRIQUE HERRERA-VIEDMA 5, (Senior Member, IEEE) 1Joint-PhD programme in Communication, Universidad de Málaga, 29071 Málaga, Spain 2Department of Computer Science and Engineering, Universidad de Cádiz, 11202 Cádiz, Spain 3Department of Audiovisual Communication and Advertising, Faculty of Communication Sciences, Universidad de Málaga, 29071 Málaga, Spain 4Department of Languages and Computer Science, Higher Technical School of Computer Engineering, Universidad de Málaga, 29071 Málaga, Spain 5Andalusian Research Institute in Data Science and Computational Intelligence, Universidad de Granada, 18071 Granada, Spain

Corresponding author: Carlos de las Heras-Pedrosa ([email protected])

This work was supported by the Programa Operativo FEDER Andalucía 2014-2020 under Grant UMA 18-FEDERJA-148.

ABSTRACT In the last decade, the advertising industry has experienced a quantum leap, powered by recent advances in neuroscience, a large investment in artificial intelligence, and a high degree of consumer expertise. Within this context, opinion mining, sentiment analysis, and emotion understanding bring us closer to one of the most sought-after objectives of advertising: to offer relevant ads at scale. The importance of studies about opinion mining, sentiment analysis, and emotion understanding in advertising has been rising exponentially over the last years. The peak of this new situation has been the interest of the research community in studying the relationship between such innovations and the spread of smart and contextual advertising. This article analyzes those works that address the relationship between sentiment analysis, opinion mining, and emotion understanding in advertising. The main objective is to clarify the current state of these studies, explore issues, methods, findings, themes, and gaps as well as to define their significance within the current convergence advertising research scenario. To reach such objectives, a bibliometric analysis was conducted, retrieving and analyzing 919 research works published between 2010 and 2019 based on results from Web of Science (WoS).

INDEX TERMS Advertising research, bibliometrics, communication, consumer behavior, emotion under- standing, opinion mining, science mapping analysis, SciMAT, sentiment analysis, VOSviewer, Web of Science (WoS).

I. INTRODUCTION The combination of technology, data, and creativity is today the driving force in the advertising landscape. Personalized advertising, ongoing data privacy concerns, shifting stake- holder power, and data-based targeting are converging to dis- rupt the future of advertising [1]. The way advertisers reach their consumers is evolving at lightning speed and consumers are moving away from traditional channels and platforms to digital media ecosystems. For an industry that requires the attention and interaction of consumers, it is essential to

The associate editor coordinating the review of this manuscript and

approving it for publication was Md. Asikuzzaman .

understand the consumer behavior and the complexities of the target audience [2].

Sentiment Analysis, opinion mining and emotion under- standing are an interdisciplinary multimodal field of study gathering between neuroscience [3]–[7], and computer sci- ence and artificial intelligence [8]–[10] that analyzes people’s attitudes, appraisals, evaluations, sentiments, opinions, and emotions to entities such as organizations, services, products, individuals, issues, events topics and their attributes [11].

The use of sentiment analysis, opinion mining and emotion understanding may be very useful as a real-time feedback loop for advertising effectiveness and might be able to predict advertising results, as well as to find and measure consumer

VOLUME 8, 2020 This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ 134563

P. Sánchez-Núñez et al.: Opinion Mining, Sentiment Analysis and Emotion Understanding in Advertising

opinions and attitudes towards their brand, products, services, and campaigns [12], [13].

A series of recent studies have indicated that analyz- ing brand sentiment provides an outstanding source of data that demonstrates the different perceptions they can qualify KPIs [14], identify influencers for the brand [15], define the brand’s reputation [16], improve the consumer experi- ence [17], determine the future of marketing strategies [18], generate leads and improve marketing campaigns and product messaging [19], [20]. Brands such as The Walt Disney Co. rely on sentiment analysis and emotion understanding to activate soundtracks when parents read stories aloud to their children or Coca-Cola, that uses opinion mining to reinvent the way consumers interact with products through smartphones [21]–[24].

Diverse industry protagonists such as media companies, digital platform businesses, agencies, advertisers, indepen- dent researchers, and consumers are aware of the urgent need to implement digital advertising transformation strategies to create a fundamentally different business that can generate sustainable profitability in the face of this disruption [25]. This consumer behavior specialization is trying to take a step further [26], leading scientists and scholars to work on responses with a high level of personalization and emotional understanding [27], fostering a sensitive connection with the consumer which is leading to increased brand recall, positive brand associations and brand awareness [28], [29].

However, some key questions and notions are still not discussed in the literature about this topic. We do not the key themes and how they evolve through the time [30], [31], pat- terns, trends, and methodologies recently used in the advertis- ing environment to respond to consumer demands. Besides, we do not know the recent scientific results on the needs of the advertising market as well as future developments in intelligent advertising [32]. In this article, we conducted a bibliometric analysis [33]–[35], a system that analyzes cita- tions and discusses scientific works published in a specific area of knowledge. Review the literature allow us to dis- cover important patterns and variables relevant to the object of study, to establish the context of the topic or problem, to synthesize and acquire a new perspective, to relate ideas and theories to their applications, to distinguish the research that has been carried out and future lines of research, to iden- tify main methodologies and research techniques as well as to place the research in a historical context to demonstrate familiarity with the latest developments [36], [37].

This bibliometric analysis aims to answer the following research questions: • RQ1. What are the key themes, incoming or outgoing topics, citation patterns, prolific authors, organizations, countries, journals, and publications detected in senti- ment analysis, opinion mining, and emotion understand- ing in the advertising ecosystems?

• RQ2. What are the thematic areas and cluster networks in sentiment analysis, opinion mining, and emotion understanding in the communication landscape?

• RQ3. What are the trends, methodologies, research gaps, and main future lines of research about the studies already carried out?

For better clarification of the results, a thematic cluster network and strategic diagrams by periods are used to catego- rize the detected topics or themes. Furthermore, we develop a performance analysis using different basic bibliometric indicators (number of received citations, number of published documents, etc.,) as well as H-Index.

This paper is organized as follows: Section 2 explains our review materials, methodology approach, and query design. In Section 3 we provide a bibliometric analysis of opinion mining, sentiment analysis and emotion understanding in advertising and Section 4 contains the discussion, conclusion and briefly accomplishes and sketches out an agenda for future research.

II. MATERIALS AND METHODS The bibliometric mapping was conducted based on scien- tific publications related to opinion mining, sentiment anal- ysis, and emotion understanding in advertising. The source of information was the Web of Science (WoS) database. The WoS, owned by Clarivate Analytics, is a collection of databases of bibliographic references and citations from peri- odicals that collect information from 1900 to the present. The choice of the WoS database was determined by the fact that it contains the most accurate and reliable research information and offers a high number of analysis tools to process it [38].

In this study, we obtained research publications indexed in WoS on opinion mining, sentiment analysis, and emotion understanding in advertising for a significantly large period of years (2010-2019), which covers almost the whole period of large scientific production in this field. Key-terms and phrases associated with emotion understanding, sentiment analysis, and opinion mining were utilized in the subject search in combination with advertising/marketing. The spe- cific search strings were formulated according to the search logic of the WoS database. Table 1 illustrates the query design, this query selects the publications according to the inclusion and exclusion criteria used, and the indexes, times- pan, and date of the data download.

To perform the review, we have used the following tools for the analysis of scientific production:

VOSviewer version 1.6.15, a software tool for constructing and visualizing bibliometric networks (including individual publications, researchers, journals); being those constructed based on co-authorship relations, co-citation, bibliographic coupling, citation and co-occurrence networks of important terms extracted from a body of scientific literature [39].

VOSviewer was used to obtain citation based-networks, analyze bibliometric networks, and create visualization maps based on network data of countries/regions, authors, organi- zations, sources, and documents. VOSviewer uses different techniques such as the network layout and network clustering (layout and clustering results can be fine-tuned using various

134564 VOLUME 8, 2020

P. Sánchez-Núñez et al.: Opinion Mining, Sentiment Analysis and Emotion Understanding in Advertising

TABLE 1. Details of dataset.

parameters) and natural language processing techniques (Rel- evant and non-relevant terms can be distinguished algorithmi- cally). VOSviewer Analysis Configuration we followed:

1) Unit of analysis: Organizations, Authors, Countries/ Regions, Sources and Documents

2) Kind of network: Citation Analysis (the relatedness of items is determined based on the number of times they cite each other)

3) Cluster network design: Network Visualization and Density Visualization (provides a quick overview of the main areas/relationships in a bibliometric network).

SciMAT version 1.1.04 (Science Mapping Analysis Soft- ware Tool), is an open-source science mapping software tool that incorporates methods, algorithms, and measures for all the steps in science mapping workflow, from preprocessing to the visualization of the results [30], [31]. SciMAT was used to study the evolution of key themes over time and the identification of developing or decreasing topics. ScIMAT Analysis Configuration we followed:

1) Unit of analysis: Words (authorRole = true, source- Role = true, addedRole = false)

2) Kind of network: Co-occurrence 3) Normalization measure: Equivalence index 4) Cluster algorithm: Centers simples 5) Max cluster size: 12 6) Min cluster size: 3 7) Evolution measure: Inclusion index 8) Overlapping measure: Jaccard index

III. RESULTS A. DISTRIBUTION OF PUBLICATIONS BY YEAR AND RECORD COUNT (2010-2019) The distribution of publications during the period 2010-2019 is shown in Figure 1. During the first lustrum (2010-2014) of the study, sustained growth of publications (n=214, 23.28%) is observed while in the second lustrum (2015-2019) it is

FIGURE 1. Distribution of publications by year and record count.

FIGURE 2. Sum of times cited by year.

detected that opinion mining, sentiment analysis and emotion understanding in advertising has suffered an exponential growth in the number of publications (n = 705, 76.71%).

B. CITATION REPORT AND RECORD COUNT In Table 2 is shown the Citation Report and the Record Count. The total publications retrieved (919) combined a sum of 7263 times cited (Figure 2), making an average of 7,9 citations per paper. The H-index is the same as 40, which means that 40 studies have received at least 40 cita- tions. The H-Index is often used to quantify an individual’s research output [40]. An extension of the H-Index to iden- tify the highly cited papers called H-Classics can be viewed in [38], and consequently, in that case, H-Classics identifies 50 highly cited papers.

VOLUME 8, 2020 134565

P. Sánchez-Núñez et al.: Opinion Mining, Sentiment Analysis and Emotion Understanding in Advertising

TABLE 2. Citation report and record count.

FIGURE 3. Document types in web of science.

C. DOCUMENT TYPES AND RECORD COUNT In Figure 3 is shown the total average % of 919 document types in Web Of Science: The largest collection of Arti- cle (485 records, 52.77%), followed by Proceedings Paper (392 records, 42.65%) Review (44 records, 4.78%), Book Chapter (28 records, 3.04%), Early Access (2 records, 0.21%) and by the end Editorial Material (1 record, 0.10%).

FIGURE 4. Research areas and record count in web of science categories.

D. WEB OF SCIENCE CATEGORIES/RESEARCH AREAS AND RECORD COUNT Figure 4 shows the Web of Science Research Areas. Among the Top 10 most representative categories in opinion mining, sentiment analysis and emotion understanding in advertis- ing we find the following: Computer Science (476 regis- ters and 51.75% of 919 works), Engineering (236 registers and 25.68% of 919 works), Business Economics (159 reg- isters and 17.30% of 919 works), Telecommunications

(62 registers and 6.74% of 919 works), Psychology (57 reg- isters and 6.20% of 919 works), Communication (41 regis- ters and 4.46% of 919 works), Social Sciences (38 registers and 4.13% of 919 works), Information Science and Library Science (31 registers and 3.37% of 919 works) and Science Technology and Other Topics (29 registers and 3.15% of 919 works).

E. PERFORMANCE INDICATORS FOR ORGANIZATIONS AND RECORD COUNT Table 3 presents the Performance Indicators for Orga- nizations and Record Count, a Top 25 selection of the most high-ranking universities along with several records; two indicators of global university ranking according to the 2019 Quacquarelli Symonds (QS) World University Rankings and 2019 Academic Ranking of World Universi- ties (ARWU) that allow us to measure the relative position in which we find the most influential institutions in senti- ment analysis, opinion mining and emotion understanding in advertising.

The relatedness of items is based on the number of times they cite each other. A minimum number of documents of an organization (5) and a minimum number of citations of an organization (5). The number of citations of an organiza- tion equals the total number of citations the documents of the organization have received in Web of Science. Of the 1146 organizations, 26 meet the threshold.

Within the first 10 universities, 40% are in the United States, followed by institutions in The Netherlands (1), Sin- gapore (1), Hong Kong (1), Denmark (1), South Korea (1) and United Kingdom (1). Further down the rankings are other institutions in China, Malaysia, Austria, Italy, The Nether- lands, Taiwan, South Korea, China, Australia, Malaysia, and Italy. The first institution in the ranking in terms of citation is the City University of Hong Kong with a total of 8 doc- uments published about sentiment analysis, opinion mining, and emotion understanding in advertising, where 8 of these studies have received 321 citations.

According to the relative position of the university rank- ing, 1st ranked is the City University of Hong Kong with a total of 8 publications and 321 citations, located within the first 201-300 (ARWU2019) and 55 (QS 2019), followed by 2nd ranked Nanyang Technological University, with a total of 10 articles published, of which 10 have been cited at least 305 times, 3rd ranked is the Copenhagen Business School, with 5 papers published and a ratio of 216 citations. Only 8 of the Top 25 university rankings are in the Top 100 ranking according to ARWU: Nanyang Technological University, University of California, San Diego, University of Florida, University of Minnesota, Aarhus University, Cornell University, University of Maryland and the National Univer- sity of Singapore.

Of these, 5 universities are in the United States while only 9 are part of the Top 100 according to QS: City Univer- sity of Hong Kong, Nanyang Technological University, Uni- versity of California San Diego, University of Nottingham,

134566 VOLUME 8, 2020

P. Sánchez-Núñez et al.: Opinion Mining, Sentiment Analysis and Emotion Understanding in Advertising

TABLE 3. Performance indicators for organizations and record count.

Korea Advanced Institute of Science and Technology, Univer- sity of Amsterdam, University of Science and Technology of China, Cornell University, University of Malaya and National University of Singapore.

F. PERFORMANCE INDICATORS FOR AUTHORS AND RECORD COUNT The Performance Indicators for Authors and Record Count can be seen in Table 4, which presents a Top 25 rank- ing of the most commanding authors in opinion mining, sentiment analysis, and emotion understanding in advertis- ing in terms of the number of documents/citations. The relatedness of items is based on the number of times they cite each other. The minimum number of documents of an author (3) and the minimum number of citations of an author (1). The number of citations of a country equals the

total number of citations the documents of the country have received in Web of Science. Of the 2853 authors, 41 meet the threshold.

G. PERFORMANCE INDICATORS FOR COUNTRIES/REGIONS AND RECORD COUNT The Performance Indicators for Countries/Regions can be seen in Table 5, which presents a Top 25 ranking of the most leading countries/regions in opinion mining, sentiment analysis, and emotion understanding in advertising in terms of the number of documents/citations. The relatedness of items is based on the number of times they cite each other (Figure 5). The minimum number of documents of a country (10) and the minimum number of citations of a country (1). The number of citations of a country equals the total number

VOLUME 8, 2020 134567

P. Sánchez-Núñez et al.: Opinion Mining, Sentiment Analysis and Emotion Understanding in Advertising

TABLE 4. Performance indicators for authors and record count.

of citations the documents of the country have received in Web of Science. Of the 88 countries, 27 meet the threshold.

The first place is occupied by the USA (206 documents and 2804 citations), followed by Italy (55 documents and 883 citations), China (117 documents and 829 citations), Australia (44 documents and 489 citations) and by the end, England (52 documents and 465 citations).

TABLE 5. Performance indicators for countries/regions.

H. PERFORMANCE INDICATORS FOR JOURNALS AND RECORD COUNT The Performance Indicators for Journals might be seen in Table 6, which presents a Top 10 ranking of the most important sources in opinion mining, sentiment analysis, and emotion understanding in advertising in terms of the number of documents/citations. The relatedness of items is based on the number of times they cite each other. The minimum num- ber of documents of a source (6) and the minimum number of citations of a source (1). The number of citations of a country equals the total number of citations the documents of the source have received in Web of Science. Of the 687 sources, 10 meet the threshold.

As shown in Table 6, the most cited journals in opinion mining, sentiment analysis, and emotion understanding in advertising have a clear focus on communication and mar- keting, artificial intelligence, computational neuroscience, or psychology among others. Being the 1st ranked Expert Systems with Applications (with 11 documents and a sum

134568 VOLUME 8, 2020

P. Sánchez-Núñez et al.: Opinion Mining, Sentiment Analysis and Emotion Understanding in Advertising

FIGURE 5. Citation analysis (countries/regions) in network visualization [39].

TABLE 6. Performance indicators for journals.

of 482 citations), 2nd ranked Decision Support Systems (8 documents and a sum of 458 citations), 3rd ranked Neu- rocomputing (with 6 documents and a sum of 112 citations), 4th ranked Computers in Human Behavior (with 8 documents and a sum of 110 citations) and 5th ranked Psychology and Marketing (with 7 documents and a sum of 79 citations) are the most cited journals in sentiment analysis, opinion mining and emotion understanding in advertising. The first, second and third are usually regarded as the three most influential sources in artificial intelligence and technology, while the fourth and fifth journals show its clear thematic connection (psychology, human behavior, and marketing).

FIGURE 6. Citation analysis (documents) in density visualization [39].

I. PERFORMANCE INDICATORS FOR PUBLICATIONS IN DENSITY VISUALIZATION Performance Indicators for Publications through Citation Analysis in item density visualization is shown in Figure 6. Relatedness of items is determined based on the number of times they cite each other (Units of analysis: documents). Of the 919 documents, 40 meets the threshold and a minimum number of citations of a document are shown (40). Some of the 49 works of the network are not connected and the largest set of connected items consists of 9 items (the figure shows all the items). In the item density visualization, items are represented by their label in a similar way as in the net- work visualization and the overlay visualization. Each point in the item density visualization has a color that indicates the density of items at that point. Only in the center of Figure 6 a recent citation network established by the authors can be appreciated: Mostafa (2013), Yu (2013), Sheng (2017), Ghose (2012), and Xu (2011).

The results reveal that there is only citation pattern con- nectivity’s in recent works where there has been a research

VOLUME 8, 2020 134569

P. Sánchez-Núñez et al.: Opinion Mining, Sentiment Analysis and Emotion Understanding in Advertising

study about social networks and text mining for consumer brand sentiment, mining comparative opinions from customer reviews for competitive intelligence, ranking systems for hotels on travel search engines by mining user-generated and crowdsourced content, the impact of social and conventional media on firm equity value (a sentiment analysis approach), and multidisciplinary perspective of big data in management research (research review).

J. PERIOD VIEW AND LONGITUDINAL RESULTS 2010-2019: STRATEGIC DIAGRAM AND PERFORMANCE MEASURES (WORDS ANALYSIS) A total of 919 web of science papers were collected, of which 916 could be used, as there were 3 that could not be included in the study due to indexing and export errors in a format not supported by SciMAT. The analysis has been divided into two consecutive lustrums (2010-2014 | 2015-2019). There are 214 documents understudy in the first subperiod and 702 documents in the second subperiod.

FIGURE 7. The strategic diagram based on Callon’s density and centrality measures.

The resulting strategic diagrams (Figure 9 and 10) shows the detected clusters of each period in a two-dimensional space and categorizes them according to their Callon’s den- sity and centrality measures. The strategic diagram is divided into 4 quadrants shown in Figure 7 (upper-right quadrant defines motor clusters, upper-left quadrant defines highly developed and isolated clusters, lower-left quadrant defines emerging or declining clusters and lower-right quadrant defines basic and transversal clusters).

FIGURE 8. Overlap fractions (incoming and outcoming keywords between successive subperiods).

Figure 8 shows the stability measures across the two con- secutive periods. The circles represent the subperiods (2) and their number of associated keywords (901 and 2592 respec- tively). The horizontal arrow represents the number of key- words shared by both periods (336) and, in parentheses,

the Similarity Index between them is shown (0.11). The upper-incoming arrow represents the number of new key- words in period 2 (565), and the upper-outcoming arrow represents the keywords that are present in period 1 but not in period 2 (2256).

1) SUBPERIOD VIEW 2010-2014: STRATEGIC DIAGRAM AND PERFORMANCE MEASURES (WORDS ANALYSIS) The subperiod (2010-2014) shows quantitative measures based on the number of documents and qualitative or impact measures based on the number of received citations/average citations of the documents and bibliometric indices such as the H-Index (Table 7 and Figure 9). We analyzed the two quadrants that we consider fundamental and most interest- ing for the development of the discipline: upper-right and lower-left.

TABLE 7. Performance measures for the themes of the subperiod 2010–2014.

FIGURE 9. Strategic diagrams for the subperiod 2010–2014 (documents count).

In the first subperiod of study (2010-2014), we witness the birth of the inter-discipline object of the study. Senti- ment analysis, opinion mining, and emotion understanding in advertising is not very developed and there are still no emerging trends or representative thematic groups.

We observe that the motor themes quadrant with the great- est number of works is focused on different clusters:

WORD-OF-MOUTH with studies in social networks text mining for consumer brand sentiments, design of rank- ing systems for hotels on travel search engines by mining

134570 VOLUME 8, 2020

P. Sánchez-Núñez et al.: Opinion Mining, Sentiment Analysis and Emotion Understanding in Advertising

user-generated and crowdsourced content, and deciphering word-of-mouth in social media through text-based metrics of consumer reviews.

VISUAL-ATTENTION cluster with studies in effects of advertising billboards during simulated driving, the influence of selective attention, and inattention to products on subse- quent choice and eye movements when viewing advertise- ments among others.

In emerging or declining themes quadrant, we find CLAS- SIFICATION cluster with studies in predicting consumer sentiments from the online text, sentiment-oriented contex- tual advertising, and emotional polarity analysis of con- sumers’ airline service tweets.

2) SUBPERIOD VIEW 2015-2019: STRATEGIC DIAGRAM, PERFORMANCE MEASURES, EMERGING OR DECLINING CLUSTERS AND MOTOR CLUSTER NETWORKS ANALYSIS (WORDS ANALYSIS) The subperiod (2015-2019) shows quantitative measures based on the number of documents and qualitative or impact measures based on the number of received citations/average citations of the documents and bibliometric indices such as the H-Index (Table 8 and Figure 10).

FIGURE 10. Strategic diagrams for the subperiod 2015–2019 (documents count).

During the second subperiod (2015-2019), we observed the fast development of research in sentiment analysis, emo- tion recognition, and opinion mining in advertising. This is the period in which the greatest scientific production exists, where we find considerably defined and powerful clusters divided into quadrants.

The results show that in motor themes quadrant (Figure 10 and Figure 11), we found a large cluster amount of research in:

MEDIA cluster with studies about Facebook as a des- tination marketing tool, eye-tracking technique to under- stand the effects of brand placement disclosure types in television programs, social media metrics, and analytics in

TABLE 8. Performance measures for the themes of the subperiod 2015–2019.

marketing-s3m and Facebook social engagement for national tourism organizations.

ONLINE-REVIEWS cluster with studies in consumer sen- timent in an online community environment, the study of the power of the "like" button and the impact in social media, social media analytics in extracting and visualizing Hilton Hotel ratings and reviews from TripAdvisor and assess- ment consumers’ satisfaction and expectations through online opinions and reviews.

CHOICE cluster with studies in first fixation and total fix- ation duration in consumer choice and visual attention toward tourism photographs with the text through an eye-tracking study.

RECALL cluster with studies in creativity, attention, and the memory for brands in outdoor advertising and effects of personalized banner ads on visual attention and recognition memory.

VOLUME 8, 2020 134571

P. Sánchez-Núñez et al.: Opinion Mining, Sentiment Analysis and Emotion Understanding in Advertising

FIGURE 11. Subperiod 2015-2019 thematic area (motor clusters).

FIGURE 12. Subperiod 2015-2019 thematic area (emerging or declining clusters).

ATTENTION cluster with studies in visual attention and responses to personalized advertising based on task cognitive demand, advertising effectiveness in travel 2.0 websites, and distraction effects of contextual advertising on online news processing through an eye-tracking study.

SENTIMENT-ANALYSIS cluster with studies in novel social media competitive analytics framework with sentiment benchmarks and business intelligence in online customer textual reviews.

PERFORMANCE cluster with studies in attention alloca- tion and memory effects when multiscreen, content compo- sition, and slot position in personalized banner ads, and how they influence visual attention in online shoppers.

In emerging or declining themes quadrant (Figure 10 and Figure 12) we found the following:

DEEP-LEARNING cluster with studies in Sitcom-star- based clothing retrieval for video advertising with deep learn- ing, inbound e-marketing using neural network-based visual and phonetic user experience analytics and automated cura- tion of brand-related social media images.

SYSTEMS cluster with studies in credibility ranking of users in big social data incorporating semantic analysis and temporal factor.

IMPACTS cluster with studies in seeking attention through an eye-tracking study of in-store merchandise displays.

CUSTOMER-SATISFACTION cluster with studies in enhancing hotel guest experience.

CLASSIFICATION cluster with studies in the analysis of geolocated Airbnb rental images in cities and predicting purchase intention according to fan page user’s sentiment.

TEXT CLASSIFICATION cluster with studies in recurrent neural networks for short text and sentiment classification and novel frameworks to detect unqualified restaurant reviews.

AUGMENTED-REALITY cluster with studies in applica- tion and scope analysis of augmented reality in marketing using image processing technique and scalable mobile image recognition for real-time video annotation among others.

We note that the CLASSIFICATION cluster has been maintained in both the first and second subperiods.

Thematic networks (Figures 11 and 12) are labeled using the name of the most significant keyword in the associated theme usually identified by the most central keyword of the theme).

K. CITATION CLASSICS IN OPINION MINING, SENTIMENT ANALYSIS AND EMOTION UNDERSTANDING IN ADVERTISING Based on the result of the query, a selection of publica- tions was performed based on the H-Classics, considered an indicator that reflects the quality of the research and its impact [38], [41], [42]. The H-Index provided by WoS encompasses all the instances of citing articles successfully linked to the cited reference. In this case, we selected the publications that have obtained at least H-Index 40 citations during the period 2010-2019, considering them the most rel- evant research publications in the field. The Citation Classics can be seen in Table 9, which presents a ranking of the most significant documents in terms of H-Classics.

The selection of studies is done according to their Rank, Title, Author, Publication Year (PY), Source Title, Special Issue (SI), Total Citation (TC), Average citation/year (AY), and Citation Timespan.

Themes in the upper-right quadrant (Motor Clusters). Themes that are well developed and important for the struc- turing of a research field (Figure 11).

134572 VOLUME 8, 2020

P. Sánchez-Núñez et al.: Opinion Mining, Sentiment Analysis and Emotion Understanding in Advertising

TABLE 9. Citation classics and relatedness of items is determined based on the number of times they are cited (publications).

Themes in the lower-right quadrant (Emerging or Declining Clusters). Themes that are important for a research field but are not developed. So, this quadrant group transversal and general, basic themes (Figure 12).

The results of 40 Times-Cited documents reveal that 12 studies have been cited at least 100 times and 34 studies that have been cited at least 50 times. The year mode is 2012 with a total of 8 publications. There are 13 of the 40 publications that correspond to scientific works published in Special Issues.

The sum of the total citation of 40 Times-Cited documents is 3408 citations and the Citation Average/Year is 11,59 cita- tions. Different journals repeated with different publications in the top 40 Times-Cited: Expert Systems with Applications (3 publications), Decision Support Systems (4 publications), Journal of Advertising (2 publications), and Food Quality and Preference (2 publications).

The 1st Ranked publication with 200 citations is the work of Mohamed M. Mostafa, a journal article published in August 2013 in Expert Systems with Applications ‘‘More than words: Social networks’ text mining for consumer brand sentiments’’. The study uses text mining techniques to inves- tigate hidden patterns in consumers’ attitudes towards global brands.

IV. DISCUSSION AND CONCLUSION Opinion mining, sentiment analysis, and emotion understand- ing are nowadays fundamental in any business development strategy, playing a big role in the advertising research ecosys- tem by helping companies to deliver tailored marketing mes- sages based on business goals, rethinking the entire strategy and personalizing the marketing messages to cater to the target audience. The complexity of the study of sentiment analysis and emotion recognition leads us to approach the study of this inter-discipline from a threefold perspective: the study of the synergies between computer vision, nat- ural language processing, and neuroscience in advertising.

This bibliometric analysis explored the factors most likely to influence how today’s advertising players will vie for relevance and market share and offers diverse visions for the future.

The bibliometric analysis allows us to highlight the follow- ing remarkable findings:

Are collected 919 publications in opinion mining, sen- timent analysis, and emotion understanding in advertising were identified in the period 2010–2019, with citation counts ranging from 317 to 4772. The results suggest that most of the research carried out is developed during the second lustrum 2015-2019 (n=705, 76.71%). The total publications retrieved combined a sum of 7263 times cited, making an average of 7,9 citations per paper. The H-index is the same as 40 (based on the study and methodology of H-Classics).

The largest collection of publications in Web of Science is Article Document Type (485 records, 52.77%) followed by Proceedings Paper Document Type (392 records, 42.65%).

The most representative Categories/Research areas in Web of Science in opinion mining, sentiment analysis and emotion understanding in advertising are the following: Computer Science (476 registers and 51.75% of 919 works), Engineer- ing (236 registers and 25.68% of 919 works) and Business Economics (159 registers and 17.30% of 919 works).

The City University of Hong Kong (China) with a total of 8 publications and 321 citations, located within the first 201-300 (ARWU2019) and 55 (QS 2019) and Nanyang Technological University (Singapore) within the 73 positions (ARWU2019) and 12 (QS 2019), with a total of 10 articles published, of which 10 have been cited at least 305 times are the main institutional contributors in the discipline in terms of citation/documents.

The most prolific authors with the highest citation impact in terms of documents/citation are Professor Cambria, from Nanyang Technological University (Singapore), and profes- sor Poria, from the Singapore University of Technology and Design (Singapore).

VOLUME 8, 2020 134573

P. Sánchez-Núñez et al.: Opinion Mining, Sentiment Analysis and Emotion Understanding in Advertising

The most important journals in terms of the number of documents/citations in opinion mining, sentiment analysis, and emotion understanding in advertising are Expert Sys- tems with Applications (11 documents and 482 citations) and Decision Support Systems (8 documents and 458 citations).

The hegemony and predominance of the USA in research on sentiment analysis and opinion mining in advertising are remarkable. Its production represents a third part of the total amount of publications with 206 documents and 2804 cita- tions. Followed by Italy with 55 documents and 883 citations and Peoples R. China with 117 documents and 829 citations. It is paradoxical how Italy is in the second position with half as many articles as China and with a similar number of citations. This denotes the scarce Italian production but of great quality and sum citation impact.

The results reveal that there is only citation pattern con- nectivity’s in recent works where there has been a research study about social networks and text mining for consumer brand sentiment, mining comparative opinions from customer reviews for competitive intelligence, ranking systems for hotels on travel search engines by mining user-generated and crowdsourced content among others.

During the first subperiod of study (2010-2014), we observed the birth of the inter-discipline object of the study. The field is not very developed at that time and there are still no emerging trends or representative thematic groups.

During the second subperiod (2015-2019) we observed the fast development of research in sentiment analysis, emo- tion recognition, and opinion mining in advertising. It is the period, the greatest scientific production happens, where we find significantly defined and powerful clusters divided into quadrants. Focusing the study on the two quadrants that we consider fundamental for the study and development of the discipline:

The motor themes focused on studies in Facebook as a des- tination marketing tool, eye-tracking technique to understand the effects of brand placement disclosure types in television programs, studies in consumer sentiment in an online com- munity environment, studies of the power of the "like" button and the impact in social media, social media analytics in extracting and visualizing ratings and reviews from TripAd- visor, assessment consumers’ satisfaction and expectations through online opinions and reviews, first fixation and total fixation duration in consumer choice, visual attention toward tourism photographs with the text through an eye-tracking study, attention and the memory for brands in outdoor adver- tising, effects of personalized banner ads on visual attention and recognition memory, content composition and slot posi- tion in personalized banner ads, and how they influence visual attention in online shoppers among others.

The emerging or declining themes are focused on studies in clothing retrieval for video advertising with a deep learning, inbound e-marketing using neural network-based visual and phonetic user experience analytics, automated curation of brand-related social media images, studies in seeking atten- tion through an eye-tracking study of in-store merchandise

displays, studies in the analysis of geolocated Airbnb rental images in cities, predicting purchase intention according to fan page users sentiment, studies in recurrent neural networks for short text and sentiment classification, novel frameworks to detect unqualified restaurant reviews and by the end, stud- ies in application and scope analysis of augmented reality in marketing using image processing technique and scal- able mobile image recognition for real-time video annotation among others.

We note that the CLASSIFICATION cluster has been maintained in both the first and second subperiods.

The findings of 40 Times-Cited documents (H-Classics) reveal that 12 studies have been cited at least 100 times and 34 studies that have been cited at least 50 times. The year 2012 was the most productive period with a total of 8 publications. The 1st Ranked publication with 200 citations is the work of Mohamed M. Mostafa, a journal article published in August 2013 in Expert Systems with Applications ‘‘More than words: Social networks’ text mining for consumer brand sentiments’’.

It is worth mentioning the practical application of the present study as it provides potentially relevant information to help understand the past, present, and future scientific structure of opinion mining, sentiment analysis, and emotion understanding in the advertising and marketing field that could help its upcoming research development.

For future research lines, it would be interesting to ana- lyze the literature through alternative metrics, explore visual features and patterns and its effects on moving ads images, develop alternative indexes to measure and analyze online reviews in electronic commerce, as well as develop ontologies that allow us to better structure the knowledge in the field of visual communication and marketing to raise new models of expert systems or decision support systems in the advertising ecosystem.

REFERENCES [1] A. Dafonte-Gómez, ‘‘The key elements of viral advertising. From motiva-

tion to emotion in the most shared videos,’’ Comunicar, vol. 22, no. 43, pp. 199–207, Jul. 2014, doi: 10.3916/C43-2014-20.

[2] Y. Yang, Y. C. Yang, B. J. Jansen, and M. Lalmas, ‘‘Computational adver- tising: A paradigm shift for advertising and marketing?’’ IEEE Intell. Syst., vol. 32, no. 3, pp. 3–6, May 2017, doi: 10.1109/MIS.2017.58.

[3] A. Baraybar-Fernández, M. Baños-González, Ó. Barquero-Pérez, R. Goya-Esteban, and A. de-la-Morena-Gómez, ‘‘Evaluation of emotional responses to television advertising through neuromarketing,’’ Comunicar, vol. 25, no. 52, pp. 19–28, Jul. 2017, doi: 10.3916/C52-2017-02.

[4] H. Gauba, P. Kumar, P. P. Roy, P. Singh, D. P. Dogra, and B. Raman, ‘‘Prediction of advertisement preference by fusing EEG response and sentiment analysis,’’ Neural Netw., vol. 92, pp. 77–88, Aug. 2017, doi: 10. 1016/j.neunet.2017.01.013.

[5] M. M. Marin, ‘‘Crossing boundaries: Toward a general model of neuroaes- thetics,’’ Frontiers Hum. Neurosci., vol. 9, pp. 1–5, Aug. 2015, doi: 10. 3389/fnhum.2015.00443.

[6] M. A. Umilta’, C. Berchio, M. Sestito, D. Freedberg, and V. Gallese, ‘‘Abstract art and cortical motor activation: An EEG study,’’ Fron- tiers Hum. Neurosci., vol. 6, pp. 1–9, Nov. 2012, doi: 10.3389/ fnhum.2012.00311.

[7] B. Xing, H. Zhang, K. Zhang, L. Zhang, X. Wu, X. Shi, S. Yu, and S. Zhang, ‘‘Exploiting EEG signals and audiovisual feature fusion for video emotion recognition,’’ IEEE Access, vol. 7, pp. 59844–59861, 2019, doi: 10.1109/ACCESS.2019.2914872.

134574 VOLUME 8, 2020

P. Sánchez-Núñez et al.: Opinion Mining, Sentiment Analysis and Emotion Understanding in Advertising

[8] R. Ji, D. Cao, Y. Zhou, and F. Chen, ‘‘Survey of visual sentiment pre- diction for social media analysis,’’ Frontiers Comput. Sci., vol. 10, no. 4, pp. 602–611, Aug. 2016, doi: 10.1007/s11704-016-5453-2.

[9] V. Campos, B. Jou, and X. Giró-i-Nieto, ‘‘From pixels to sentiment: Fine- tuning CNNs for visual sentiment prediction,’’ ImageVis.Comput., vol. 65, pp. 15–22, Sep. 2017, doi: 10.1016/j.imavis.2017.01.011.

[10] Z. Jianqiang, G. Xiaolin, and Z. Xuejun, ‘‘Deep convolution neu- ral networks for Twitter sentiment analysis,’’ IEEE Access, vol. 6, pp. 23253–23260, 2018, doi: 10.1109/ACCESS.2017.2776930.

[11] B. Pang and L. Lee, ‘‘Opinion mining and sentiment analysis,’’ Found. Trends Inf. Retr., vol. 2, nos. 1–2, pp. 1–135, 2008, doi: 10.1561/ 1500000011.

[12] J. S. Pedro, T. Yeh, and N. Oliver, ‘‘Leveraging user comments for aesthetic aware image search reranking,’’ in Proc. 21st Annu. Conf. World Wide Web (WWW), 2012, pp. 439–448, doi: 10.1145/2187836.2187896.

[13] C. De Las Heras-Pedrosa, D. Rando-Cueto, C. Jambrino-Maldonado, and F. J. Paniagua-Rojano, ‘‘Analysis and study of hospital communication via social media from the patient perspective,’’ CogentSocialSci., vol. 6, no. 1, Jan. 2020, doi: 10.1080/23311886.2020.1718578.

[14] W. Wereda and J. Woźniak, ‘‘Building relationships with customer 4.0 in the era of marketing 4.0: The case study of innovative enterprises in Poland,’’ Social Sci., vol. 8, no. 6, p. 177, Jun. 2019, doi: 10.3390/ socsci8060177.

[15] M. Soleymani, D. Garcia, B. Jou, B. Schuller, S.-F. Chang, and M. Pantic, ‘‘A survey of multimodal sentiment analysis,’’ Image Vis. Comput., vol. 65, pp. 3–14, Sep. 2017, doi: 10.1016/j.imavis.2017.08.003.

[16] H. Kim, W.-M. Hur, and J. Yeo, ‘‘Corporate brand trust as a mediator in the relationship between consumer perception of CSR, corporate hypocrisy, and corporate reputation,’’ Sustainability, vol. 7, no. 4, pp. 3683–3694, Mar. 2015, doi: 10.3390/su7043683.

[17] M. Romero-Calmache and C. Fanjul-Peyró, ‘‘Advertising in the digital age: The microsite as a strategic factor in on-line advertising campaigns,’’ Comunicar, vol. 17, no. 34, pp. 125–134, Mar. 2010, doi: 10.3916/C34- 2010-03-12.

[18] J. I. Peláez, E. A. Martínez, and L. G. Vargas, ‘‘Products and services val- uation through unsolicited information from social media,’’ Soft Comput., vol. 24, no. 3, pp. 1775–1788, Apr. 2019, doi: 10.1007/s00500-019-04005- 3.

[19] Y. Yuan, F. Wang, J. Li, and R. Qin, ‘‘A survey on real time bidding advertising,’’ in Proc. IEEE Int. Conf. Service Oper. Logistics, Inform. (SOLI), Oct. 2014, pp. 418–423, doi: 10.1109/SOLI.2014.6960761.

[20] C. L. Philip Chen and C.-Y. Zhang, ‘‘Data-intensive applications, challenges, techniques and technologies: A survey on big data,’’ Inf. Sci., vol. 275, pp. 314–347, Aug. 2014, doi: 10.1016/j.ins.2014.01. 015.

[21] S. Druga, R. Williams, C. Breazeal, and M. Resnick, ‘‘‘Hey Google is it ok if I eat you?’: Initial explorations in child-agent interaction,’’ in Proc. Conf. Interact. Design Children, Jun. 2017, pp. 595–600, doi: 10.1145/ 3078072.3084330.

[22] V.-D. Păvăloaia, E.-M. Teodor, D. Fotache, and M. Danileţ, ‘‘Opinion mining on social media data: Sentiment analysis of user preferences,’’ Sus- tainability, vol. 11, no. 16, p. 4459, Aug. 2019, doi: 10.3390/su11164459.

[23] O. Gruebner, S. R. Lowe, M. Sykora, K. Shankardass, S. V. Subrama- nian, and S. Galea, ‘‘A novel surveillance approach for disaster men- tal health,’’ PLoS ONE, vol. 12, no. 7, pp. 1–15, 2017, doi: 10.1371/ journal.pone.0181233.

[24] A. Pettit, ‘‘Identifying the real differences of opinion in social media sentiment,’’ Int. J. Market Res., vol. 55, no. 6, pp. 757–767, Nov. 2013, doi: 10.2501/ijmr-2013-065.

[25] J. R. Saura, P. Palos-Sanchez, and B. Rodríguez Herráez, ‘‘Digital market- ing for sustainable growth: Business models and online campaigns using sustainable strategies,’’ Sustainability, vol. 12, no. 3, p. 1003, Jan. 2020, doi: 10.3390/su12031003.

[26] K. Z. K. Zhang and M. Benyoucef, ‘‘Consumer behavior in social com- merce: A literature review,’’ Decis. Support Syst., vol. 86, pp. 95–108, Jun. 2016, doi: 10.1016/j.dss.2016.04.001.

[27] E. Añaños-Carrasco, ‘‘Eyetracker technology in elderly people: How inte- grated television content is paid attention to and processed,’’ Comunicar, vol. 23, no. 45, pp. 75–83, Jul. 2015, doi: 10.3916/C45-2015-08.

[28] D. Borth, T. Chen, R. Ji, and S.-F. Chang, ‘‘SentiBank: Large-scale ontol- ogy and classifiers for detecting sentiment and emotions in visual con- tent,’’ in Proc. 21st ACM Int. Conf. Multimedia (MM), 2013, pp. 459–460, doi: 10.1145/2502081.2502268.

[29] A. S. Baron, G. Zaltman, and J. Olson, ‘‘Barriers to advancing the science and practice of marketing,’’ J. Marketing Manage., vol. 33, nos. 11–12, pp. 893–908, Jul. 2017, doi: 10.1080/0267257X.2017. 1323839.

[30] M. J. Cobo, A. G. López-Herrera, E. Herrera-Viedma, and F. Herrera, ‘‘An approach for detecting, quantifying, and visualizing the evolu- tion of a research field: A practical application to the fuzzy sets theory field,’’ J. Informetrics, vol. 5, no. 1, pp. 146–166, Jan. 2011, doi: 10.1016/j.joi.2010.10.002.

[31] M. J. Cobo, A. G. López-Herrera, E. Herrera-Viedma, and F. Herrera, ‘‘SciMAT: A new science mapping analysis software tool,’’ J. Amer. Soc. Inf. Sci. Technol., vol. 63, no. 8, pp. 1609–1630, Aug. 2012, doi: 10.1002/asi.22688.

[32] P. Sánchez-Núñez, C. de las Heras-Pedrosa, and J. I. Peláez, ‘‘Opinion mining and sentiment analysis in marketing communications: A science mapping analysis in Web of science (1998–2018),’’ Social Sci., vol. 9, no. 3, p. 23, Feb. 2020, doi: 10.3390/socsci9030023.

[33] A. Bolderston, ‘‘Writing an effective literature review,’’ J. Med. Imag. Radiat. Sci., vol. 39, no. 2, pp. 86–92, Jun. 2008, doi: 10.1016/j. jmir.2008.04.009.

[34] J. Dumay, C. Bernardi, J. Guthrie, and P. Demartini, ‘‘Integrated report- ing: A structured literature review,’’ Accounting Forum, vol. 40, no. 3, pp. 166–185, Sep. 2016, doi: 10.1016/j.accfor.2016.06.001.

[35] J. Montero-Díaz, M.-J. Cobo, M. Gutiérrez-Salcedo, F. Segado-Boj, and E. Herrera-Viedma, ‘‘A science mapping analysis of ‘communica- tion’ WoS subject category (1980-2013),’’ Comunicar, vol. 26, no. 55, pp. 81–91, Apr. 2018, doi: 10.3916/C55-2018-08.

[36] J. Webster and R. T. Watson, ‘‘Analyzing the past to prepare for the future: Writing a literature review,’’ MIS Quart., vol. 26, no. 2, pp. 13–23, 2002. [Online]. Available: https://www.jstor.org/stable/4132319?seq=1

[37] W. Glanzel, ‘‘National characteristics in international scientific co- authorship relations,’’ Scientometrics, vol. 51, no. 1, pp. 69–115, 2001, doi: 10.1023/A:1010512628145.

[38] M. A. Martínez, M. Herrera, E. Contreras, A. Ruíz, and E. Herrera-Viedma, ‘‘Characterizing highly cited papers in social work through H-classics,’’ Scientometrics, vol. 102, no. 2, pp. 1713–1729, Feb. 2015, doi: 10.1007/s11192-014-1460-y.

[39] N. J. van Eck and L. Waltman, ‘‘Software survey: VOSviewer, a com- puter program for bibliometric mapping,’’ Scientometrics, vol. 84, no. 2, pp. 523–538, Aug. 2010, doi: 10.1007/s11192-009-0146-3.

[40] J. E. Hirsch, ‘‘An index to quantify an individual’s scientific research output,’’ Proc. Nat. Acad. Sci. USA, vol. 102, no. 46, pp. 16569–16572, Nov. 2005, doi: 10.1073/pnas.0507655102.

PABLO SÁNCHEZ-NÚÑEZ was born in 1993. He received the bachelor’s degree in design (graph- ics) from Universitat Ramon Llull, Barcelona, Spain, in 2015, and the master’s degree in com- mercial and marketing management from the Uni- versidad a Distancia de Madrid, Madrid, Spain, in 2017. He is currently pursuing the Ph.D. degree in the Joint-Doctorate Programme in Communica- tion (Advertising and Public Relations) with the Universidad de Cádiz, the Universidad de Huelva,

the Universidad de Málaga, and the Universidad de Sevilla. From 2014 to 2015, he coursed an Erasmus + Scholarship (European Commission) at Mimar Sinan Güzel Sanatlar Üniversitesi, Istanbul, Turkey. He was a Visiting Research Student with the Institute for Digital Communications (IDCOM), The University of Edinburgh, Edinburgh, U.K., in 2019. His research interests include science mapping analysis, bibliometrics, scientometrics, opinion mining, sentiment analysis, emotion understanding, data visualiza- tion, marketing, graphic design, corporate communication, and scientific communication.

VOLUME 8, 2020 134575

P. Sánchez-Núñez et al.: Opinion Mining, Sentiment Analysis and Emotion Understanding in Advertising

MANUEL J. COBO was born in 1982. He received the M.Sc. and Ph.D. degrees in computer sciences from the University of Granada, Granada, Spain, in 2008 and 2011, respectively. He is currently an Associate Professor with the Department of Computer Science and Engineering, Universidad de Cádiz, Algeciras, Spain. His research interests include science mapping analysis, bibliometrics, scientometrics, text mining, graph mining, data visualization, quality evaluation, decision making, and recommender systems.

CARLOS DE LAS HERAS-PEDROSA received the bachelor’s degree in business and economics and the Ph.D. degree in advertising and public relations. He began his career path, in 1989, at the multinational Fiat Auto España S.A., where he performed his duties as a Regional Logistics Man- ager. Later, in 1991, he is hired by multinational D’Arcy Masius Benton & Bowles as a Regional Promotion and Advertising Manager responsible for the account’s management of Fiat Auto España

S.A. The scope of application of his responsibility fell within the regions of Andalusia and Valencia. Among his academic and management positions, he has been the Co-Director of the Ph.D. Program in Organizational Com- munication, the Director of the Master’s Program in Communication and Tourism Management, and the Vice Dean of the School of Communication. Furthermore, he has exercised his duties as an Academic Coordinator of the Exchange and Mobility Program (PIMA) at the University of Málaga and the General Assistant Director of the Communication, Protocol, and Presidents’ Cabinet. For eight years, he has held the position of Vice-President for the Institutional Relations and President’s Cabinet at the University of Málaga. He has been a Visiting Researcher with the University of Miami, USA, The University of Sheffield, the University of Cardiff, the University of Leeds, U.K., and the University of Furtwangen (GER). He has also been a Visiting Professor of the Doctoral Program at the Universidad de Guadala- jara, México, the Universidad de Barinas, Universidad del Zulia, and the Universidad de Oriente, Venezuela. He is currently a Professor in audiovisual communication and advertising at University of Málaga. He also teaches a course on institutional communication. He is also the Coordinator of the Postgraduate Program Strategic Management and Innovation in Commu- nication. His research interests include health communication, institutional and political communication, and tourism communication. The latter further develops aspects related to fundraising, as well as the image of institutions and their leaders.

JOSÉ IGNACIO PELÁEZ received the degree in computer science from the University of Granada and the Ph.D. degree in computer science from the University of Granada in 2000. He is currently a Professor with the Department of Languages and Computer Sciences, Universidad de Málaga, and the Director of the Metric and Intangibles Management Chair. He has published more than 45 articles in refereed journals in the fields of majority operator, metaheuristics to design com-

posites, consistency, and decision making in economics and government. His research interests include intangibles, corporate reputation, consistency, aggregation operator, and business intelligence systems.

ENRIQUE HERRERA-VIEDMA (Senior Mem- ber, IEEE) received the M.Sc. and Ph.D. degrees in computer science from the Universidad de Granada, Granada, Spain, in 1993 and 1996, respectively. He is currently a Professor of com- puter science and the Vice-President for research and knowledge transfer with the Universidad de Granada. His current research interests include group decision making, consensus models, lin- guistic modeling, aggregation of information,

information retrieval, bibliometric, digital libraries, web quality evaluation, recommender systems, and social media. His H-index is 71 with more than 21 000 citations received in Web of Science and 91 in Google Scholar with more than 31 000 cites received. He has been identified as one of the world’s most influential researchers by the Shanghai Center and Thomson Reuters/Clarivate Analytics in both Computer Science and Engineering, from 2014 to 2019. He is the Vice President for Publications in IEEE SMC Society and an Associate Editor for several journals, such as the IEEE TRANSACTIONS ON FUZZY SYSTEMS, the IEEE TRANSACTIONS ON SYSTEMS, MAN, AND CYBERNETICS: SYSTEMS, the IEEE TRANSACTIONS ON INTELLIGENT TRANSPORT SYSTEM, Information Sciences, Applied Soft Computing, Soft Computing, Fuzzy Optimization and Decision Making, and Knowledge-Based Systems.

134576 VOLUME 8, 2020