Global research on ubiquitous learning: A network and output approach

Ubiquitous learning refers to the advancement of online learning, adapted to the development of communication and informatics. The study aimed to carry out a bibliometric approach to ubiquitous learning worldwide indexed in Scopus from 2003 to 2023. Methodologically, it was a quantitative, bibliometric study. Scientific output indicators were generated from 2668 selected Scopus papers using English keywords ("ubiquitous" AND "learning"). From 2007 to 2023, the number of published documents increased (90.9%), indicating a growing interest in the subject of study. The United States has the highest scientific production (21%), and the CNRS Centre National de la Recherche Scientifique has the most publications (n=39). The journal IEEE Access received 2514 citations, with the author Cook, D.J. (n=532) being the most cited. It is concluded that the orientation on ubiquitous learning is dynamic and increasingly linked to the development of information and communication technologies (ICT), as this learning style uses technology to give students more independence over their learning process.


Introduction
The advent of the Internet and other forms of digital technology in recent decades has led to significant social changes that have contributed to shaping a new cultural paradigm. 1Consequently, the digital era can be defined as a technological civilization in which the continuous and irreversible advancement of digital technology and the Internet has altered human behavior and organizational and productive paradigms. 2n that order of ideas, a new way of life in a networked society has emerged due to technological advances. 3][6] On the other hand, the advantage of mobile devices over previous technologies for learning is important.Thanks to their connectivity, they can be used in formal and informal educational contexts, alone or in collaboration with others. 7ue to this portability, new pedagogical approaches and lines of research have emerged, such as ubiquitous learning (u-learning). 8,9Firstly, ubiquity in learning is recognized when the same student can be virtually present in different places simultaneously. 10,114][35] Additionally, scientific data collected and preserved in databases are analyzed through meta-analysis to reveal patterns, relationships, trends, and indications in this field and to provide reliable data about the people and procedures involved in scientific discovery. 36,37onsidering the changes that occur due to the emphasis on localized and contextualized learning, shortly, teachers and students are expected to be present across multiple digital platforms. 38,39Therefore, it is important to explain and represent the academic community's understanding of trends in ubiquitous learning by categorizing information according to the year of publication, country, thematic area, document type, source, and authorship.Based on this premise, the study aims to conduct a bibliometric approach to ubiquitous learning worldwide indexed in Scopus from 2003 to 2023.

Methods
A bibliometric study was conducted on scientific papers published worldwide on ubiquitous learning in the last twenty years (2003-2023).1][42][43][44] Additionally, the information used was extracted from Elsevier's Scopus database.
Regarding the information selection process, an exhaustive search was performed on the abstract, title, or keywords of articles using the search terms "ubiquitous" AND "learning". 28,45Subsequently, the resulting metadata was extracted after applying filters based on various descriptors, such as publication year (2003-2023) and document type.In this way, a subset of 4020 documents was obtained, from which a sample of 2668 documents was extracted for analysis after duplicate exclusion and metadata normalization.Finally, a co-occurrence analysis was conducted, examining the relationships between keywords and other general categories of scientific production.In this regard, the focus is on the quantitative analysis of bibliometric data, 46 providing details such as publication year, source or journal, country of origin, author, document type, thematic area, and institutional affiliation of the author, all related to ubiquitous learning at a global level.Descriptive statistics were also collected and analyzed in Excel, and the co-occurrence map data was visualized using VOSviewer V_1.6.19.

Results
For the development of the bibliometric analysis, 2668 documents from Scopus on ubiquitous learning published between 2003 and 2023 were chosen.As shown in Figure 1, from 2007 to 2022, the scientific community had exponential growth in producing 2424 documents, representing 90.9%.Likewise, the years with the highest publication were 2022 (n=455), 2021 (n=391), and 2020 (n=303), which together represent 43.1% of all worldwide publications during the chosen period.

Figure 1. Documents published by year
Source: Scopus data (2023) Table 1 summarizes the contributions of 105 countries in scientific production, highlighting the regions where the largest number of studies on the subject have been published.The United States ranks first with 21% (n=850) of the published papers, followed by the United Kingdom (10%) and China (10%).In addition, the predominant language of publication in 96% was English, while only 4% were in Spanish and Portuguese.  2 shows the top journals regarding the number of articles on the topic studied.With 169 articles, Lecture Notes in Computer Science outnumbered most of the journals, compared to IEEE Access (n=109), Plos One (n=49), Sensors Switzerland (n=47), and Sensors (n=43).In addition, there is a preponderance of journals published in the United States and the United Kingdom, which have a high impact factor, and the vast majority are in the top two quartiles.Source: Scopus data (2023) From the data collected from the selected sources, a bibliographic clustering analysis was performed to establish the sets of sources or journals (Figure 3).Consequently, six different main clusters were identified: a first focus was the journal IEEE Access with 2514 citations, followed by Plos One (1825 citations), Lecture Notes in Computer Science (1688 citations), Sensors Switzerland (1208 citations), Proceedings of the National Academy of Sciences of the United States of America (1078 citations) and Journal of Neuroscience (777 citations).That is, the bibliographic clustering analysis showed a strong correlation between sources with many citations in common with these journals.For their part, academics from 160 institutions contributed to 2668 papers.Figure 4 shows that, over the chosen study period, most of the papers on ubiquitous learning were published by the CNRS Centre National de la Recherche Scientifique (n=39), followed by University College London (n=36) and Cambridge (n=36).In contrast, Imperial College London has 32 developed papers, compared to the Chinese Academy of Sciences, with 27 published scientific papers.The publications were written by 160 authors from 160 different academic institutions.Table 3 shows that Cook, D.J. is the author who has written the most scientific papers (n=9).He is followed in several publications by Asensio-Pérez, J.I. (n=8), Muñoz-Cristóbal, J.A. (n=7), and finally Barbosa, J.L.V. and Martínez-Monés, A. with six publications each.To complement the analysis of publications by the author, Figure 5 presents the three most cited authors with papers on ubiquitous learning: Cook, D.J., with 532 citations.While in second place, Roy, K. and Agrawal, A. stand out, each with 286 citations in published papers.Likewise, as shown in Figure 6, computer science represents 25% of scientific production, engineering 13%, and social sciences 10%.On the other hand, when analyzing the production according to the different types of documents created, it can be seen that scientific articles represent the largest part of the production (88%), followed by conference-paper (9%).In comparison, book chapters and books only represent 2%. Figure 6 shows the study results on keyword co-occurrence by selecting terms with more than three occurrences in the title, keyword, and abstract fields.The degrees of connectedness estimated by VOSviewer were used to designate each color to the set of words related to each other.
• Blue cluster."ubiquitous learning" (n=118 occurrences) is associated with the words: students, mobile learning, teaching, higher education, human-computer interaction, and internet.
• Red cluster."pattern recognition" (n=85 occurrences) clusters the following words: big data, data mining, intelligent systems, wearable sensors, smartphones, forecasting, deep neural networks, and convolutional neural networks.The clusters highlight that the most commonly used terms are those that are intrinsically related to the topic of the study.

Discussion
The results demonstrate a growing trend in the number of scientific articles addressing the topic of ubiquitous learning each year, with the highest total number of publications in 2020, 2021, and 2022 (n=1149; 43.1%).According to Berrones et al. (2023), 21 there has been an increase in the volume of publications on the topic in recent years.This is due to the widespread adoption of ICT in classrooms, enabling significant changes in pedagogy.][49] On the other hand, the most relevant institution in terms of the number of publications was the CNRS Centre National de la Recherche Scientifique, with thirty-nine publications.Additionally, the United States ranks first in scientific production with 21%, with English as the predominant language in most publications (96%).Furthermore, the author with the most publications (n=9) was Cook, D.J. from Washington State University Pullman.It was also evident that most articles were published in Lecture Notes In Computer Science (n=169) and IEEE Access (n=109).However, IEEE Access, Plos One, Lecture Notes in Computer Science, and Sensors Switzerland were the journals with the highest impact.Academic progress is impossible without international cooperation. 50New paradigms for the design, use, implementation, and evaluation of educational technologies emerge as their incorporation in the classroom is considered. 51,524][55] This helps researchers tackle the challenge of scientific production and define ubiquitous learning as learning that incorporates the elements above and all the technological tools available today for instruction and learning. 56nterdisciplinarity is also observed in fields such as engineering, social sciences, mathematics, and others, but the documents studied here stand out in computer science.Additionally, scientific articles accounted for 88% of the total document production.The most co-occurring keyword was "ubiquitous learning," which forms the basis of the study.][59] Therefore, the purpose of ubiquitous learning is to improve students' performance and expand their capabilities in their learning process through technologymediated contextual teaching.Lastly, the keywords surrounding ubiquitous learning point towards a multidisciplinary strategy, as demonstrated by its applications in computer science, education, and mathematics. 14,60,61[64]

Conclusions
According to the purpose of the study, based on the bibliometric approach to ubiquitous learning literature worldwide indexed in the Scopus database between 2003 and 2023, an exponential trend is evident from 2007 to 2022, representing 90.9% (n=2424) of the published works in the studied period.Additionally, out of the entire scientific production from 105 countries, 21% (830 documents) can be attributed to the United States.Furthermore, Lecture Notes In Computer Science has published more articles than any other journal (n=169), while the CNRS Centre National de la Recherche Scientifique has produced 39 academic works.Cook, D.J. is the author with the most scientific publications on ubiquitous learning (n=9) and citations in their works (532 citations).It was also observed that the majority of the works were scientific articles (88%) in the field of computer science (25%), engineering (13%), and social sciences (10%).Moreover, based on the keyword analysis results conducted using the VOSviewer program, "ubiquitous learning" was the most frequently used term (n=118 occurrences).Based on the analysis of the 2668 documents comprising the study sample, it is concluded that the research orientation on ubiquitous learning is dynamic and increasingly linked to the development of information and communication technology (ICT), as this learning style utilizes technology as a tool for students to access the information they need at any time, providing them with greater independence in their learning process.

Source:
Scopus data (2023) On the other hand, Figure 2 shows the trend of global articles published in the five most prominent journals on ubiquitous learning.In particular, the journal Lecture Notes in Computer Science records a steady publication of papers from 2004 to the present, with 2020 having the most published papers (n=16).Similarly, between 2017 and 2023, 109 scientific papers were published in the IEEE Access Journal, with 2019 being the year with the most scientific production (n=32).

Figure 2 .
Figure 2. Publication of papers in the main sources.

Figure 3 .
Figure 3. Clustering density map of sources or journals.

Figure 4 .
Figure 4. Papers published by institution

Figure 5 .
Figure 5.Most cited papers by the author

Table 4 .
Publication of documents by subject area and type of publication