c e p s Journal Analysis of Social Networks and Video Conferencing Systems in the Educational Context Through Data Science Ricardo-Adán Salas-Rueda* 1 , Erika-Patricia Salas-Rueda 2 and Rodrigo-David Salas-Rueda 3 • The aim of this mixed research is to analyse the perceptions of univer - sity students about the use of social networks and video conferencing systems during the COVID-19 post-pandemic through data science. The participants are 103 students of the Faculty of Sciences at the National Autonomous University of Mexico. The results of the deep learning al - gorithm indicate that social networks and video conferencing systems positively impact student autonomy and the exchange of ideas. The random forest algorithm facilitated the creation of the models on these tools considering the characteristics of the participants. Social networks facilitate use of multimedia resources, publication of school content and review of information. Likewise, video conferencing systems facilitate the realisation of classes in virtual modality through recordings and interaction between the educator and students. In conclusion, the use of social networks and video conferencing systems favour the planning and execution of new school activities at home and in the classroom. Keywords: social networks, video conferencing systems, data science, deep learning 1 *Corresponding Author. National School of Earth Sciences, National Autonomous University of Mexico, Mexico; adansalas@comunidad.unam.mx. 2 Monterrey Institute of Technology and Higher Studies, Mexico. 3 Metropolitan Autonomous University, Mexico. DOI: https://doi.org/10.26529/cepsj.2059 Received: 15 December 2024, Accepted: 30 June 2025, Published on-line as Recently Accepted Paper: July 2025 analysis of social networks and video conferencing systems in the educational ... 2 Analiza družbenih omrežij in videokonferenčnih sistemov v izobraževalnem kontekstu s pomočjo znanosti o podatkih Ricardo-Adán Salas-Rueda, Erika-Patricia Salas-Rueda in Rodrigo-David Salas-Rueda • Cilj te mešane raziskave je analizirati zaznave visokošolskih študentov o uporabi družbenih omrežij in videokonferenčnih sistemov v času po pandemiji covida-19 s pomočjo znanosti o podatkih. Udeleženci so bili 103 študentje Fakultete za znanosti na Nacionalni avtonomni univer - zi v Mehiki. Izsledki algoritma globokega učenja kažejo, da družbena omrežja in videokonferenčni sistemi pozitivno vplivajo na avtonomi - jo študentov in izmenjavo idej. Algoritem naključnega gozda je olajšal ustvarjanje modelov za ta orodja, upoštevajoč značilnosti udeležencev. Družbena omrežja olajšujejo uporabo multimedijskih virov, objavo šol - skih vsebin in pregled informacij. Podobno videokonferenčni sistemi olajšujejo izvajanje pouka v virtualni obliki s pomočjo posnetkov ter in - terakcije med učiteljem in študenti. Sklepno lahko rečemo, da uporaba družbenih omrežij in videokonferenčnih sistemov spodbuja načrtova - nje in izvajanje novih šolskih dejavnosti doma in v učilnici. Ključne besede: družbena omrežja, videokonferenčni sistemi, znanost o podatkih, globoko učenje c e p s Journal 3 Introduction In order to face the challenge of the SARS-CoV-2 virus, schools, colleges and universities offered courses with the support of information and communi - cation technologies (ICT s) (Al-Balushi et al., 2022; Isman et al., 2023; Kusuman - ingdyah et al., 2024; Rice, 2022). Social networks were used by teachers to offer new learning environments where the students actively participate in the edu - cational activities (Al-Balushi et al., 2022; Caratiquit & Caratiquit, 2023; Guil - lén-Gámez et al., 2022). Video conferencing systems allow the realisation of the educational process at any time, as teachers can present topics from anywhere (Bailey, 2022; Nti et al., 2022; Unal & Yilmaz, 2024; Walcott-Bedeau, 2022). During the COVID-19 pandemic, the incorporation of technological tools caused educators to modify their school activities and teaching methods (Camilleri & Camilleri, 2022; Cvitković et al., 2024; Rice, 2022; Sahin-Dogruer, 2023). The Internet facilitated the use of digital tools such as social networks and video conferencing systems in the teaching-learning process (Camilleri & Camilleri, 2022; Kastiro et al., 2022; Lena-Acebo et al., 2023). Educators use social networks to share multimedia resources and dis - seminate course information (Bendayan et al., 2024; Caratiquit & Caratiquit, 2023; Lundgren et al., 2022; Muls et al., 2020). These tools favour the organisa - tion of discussion forums and communication in the mixed modality (Caratiq - uit & Caratiquit, 2023; Lundgren et al., 2022; Quesnelle & Montemayor, 2020). Video conferencing systems are essential technological tools for com - munication in distance education (Bailey, 2022; Faner et al., 2022; Walcott-Be - deau, 2022). Educators use Google Meet, Microsoft T eams and Zoom to present topics remotely (Camilleri & Camilleri, 2022; Nguyen et al., 2021). Social networks and video conferencing systems allow the construction of creative educational spaces where multimedia resources are shared during the learning process (Chen, 2022; Nti et al., 2022; Walcott-Bedeau, 2022), while data science and machine learning algorithms allow users to find new informa - tion that facilitates the understanding of educational phenomena related to the integration of tools both inside and outside the classroom (Chen, 2022; Koyun - cu et al., 2022; Nti et al., 2022; Žabkar et al., 2023). Social networks Social networks are used as educational tools because they facilitate in - teraction, communication and collaboration between participants (Al-Balushi et al., 2022; Lundgren et al., 2022; Muls et al., 2020). Together with teachers, analysis of social networks and video conferencing systems in the educational ... 4 educational institutions have transformed the learning process in medicine and science through social networks (García-García et al., 2023; Hasiloglu et al., 2020; Lundgren et al., 2022). During the COVID pandemic, the incorporation of social networks in courses increased due to the ease of use and availability of these technological tools (Al-Balushi et al., 2022; Lundgren et al., 2022; Muls et al., 2020). Face - book facilitated the realisation of collaborative activities through the exchange of comments on the wall (Muls et al., 2020). Social networks have modified interaction between the educator and students (García-García et al., 2023; Hasiloglu et al., 2020; Lundgren et al., 2022). In science courses, the use of Facebook and Twitter favours debate (Lun - dgren et al., 2022), while teachers also use social networks to promote the de - velopment of students through the dissemination of information and multime - dia resources concerning science (Hasiloglu et al., 2020). The use of Facebook in medical courses fosters debate and participa - tion on the Internet (García-García et al., 2023; Nti et al., 2022; Quesnelle & Montemayor, 2020). Finally, the incorporation of social networks in universi - ties favours the creation of spaces for reflection and discussion (Bendayan et al., 2024; García-García et al., 2023; Guillén-Gámez et al., 2022; Nti et al., 2022). Video conferencing systems Learning management and video conferencing systems such as Google Meet, Microsoft Teams and Zoom enable the organisation of creative activities and the realisation of courses remotely (Camilleri & Camilleri, 2022; Nguyen et al., 2021; Walcott-Bedeau, 2022). For example, video conferencing systems such as Zoom facilitated the understanding of topics on medicine, English language and science, as students actively participated and resolved their doubts (Bailey, 2022; Faner et al., 2022; Walcott-Bedeau, 2022). In Foreign Language courses, Zoom has been shown to facilitate com - munication, understanding of topics and analysis of English language content (Bailey, 2022). The video conferencing system studied even increased academic performance in the distance modality (Bailey, 2022). In another study, medical students used Zoom to assimilate knowledge about biochemistry and interact with the teacher in real time (Faner et al., 2022). This video conferencing system allows the participants in the educational process to acquire a central role (Bailey, 2022; Faner et al., 2022; Walcott-Bedeau, 2022). Zoom also facilitated communication between participants during a preclinical science course (Walcott-Bedeau, 2022). Video conferencing systems c e p s Journal 5 allow the resolution of doubts in real time, the presentation of school topics from anywhere, participation from anywhere and the execution of the learning process in the distance modality (Bailey, 2022; Faner et al., 2022; Unal & Yilmaz, 2024; Walcott-Bedeau, 2022). Data science in the educational field Data science and machine learning algorithms have been used to under - stand the educational phenomena related to learning, academic performance, school dropout, motivation and the incorporation of technology in school ac - tivities (Hussain & Khan, 2023; Lincke et al., 2021; Sghir et al., 2023). According to Lincke et al. (2021), machine learning algorithms are used to personalise course information, analyse the impact of technological tools on educational institutions and adapt the content of educational platforms to the needs of students. In fact, machine learning algorithms such as k-nearest neighbors, linear regression, Bayesian classifier, decision tree, random forest, support vector ma - chine and deep learning allow forecasting events concerning the educational process and technological advances (Lincke et al., 2021; Sghir et al., 2023). According to Li et al. (2025), the deep learning algorithm uses neural networks that facilitate finding the most efficient predictive models in decision making. This machine learning algorithm divides the sample into two sections: the training section to find the forecast function, and the evaluation section to determine the most efficient predictive model by obtaining the smallest squared error (Li et al., 2025; Shiao et al., 2023). The deep learning algorithm is used to predict events related to the teaching-learning process, as this artificial intelligence technique establishes precise and reliable forecast models through hyperparameters such as the size of the training and evaluation sections, hidden layers and activation (Li et al., 2025; Shiao et al., 2023). Salas-Rueda et al. (2025a) used the deep learning algorithm with Tanh activation and various sizes of the training and evaluation sections to establish predictive models on educational aspects such as active role, communication and comprehension of topics in a mathematics course and an information and communication technologies course. Iyamuremye et al. (2024) and Ayanwale et al. (2024) highlight the im- (2024) highlight the im - portance of machine learning algorithms in the field of educational research due to the increase of articles published in this area; for example, the random forest algorithm is used to create tree-shaped predictive models (Beseiso, 2025; analysis of social networks and video conferencing systems in the educational ... 6 Jin, 2025). This artificial intelligence technique builds various trees until it finds the most efficient predictive model considering the accuracy aspect (Beseiso, 2025; Jin, 2025). Salas-Rueda et al. (2025b) used the random forest algorithm to create predictive models of enthusiasm and motivation during the use of an educa - tional web application on mathematics in an applied geography degree pro - gramme. Similarly, Jin (2025) used this artificial intelligence technique to pre - dict academic performance by considering teacher characteristics, educational resources, pedagogical strategies and parental interaction. Lincke et al. (2021) used the algorithms of logistic regression, linear re - gression, extreme gradient-boosted tree, decision tree, deep learning and ran - dom forest to determine the most appropriate multimedia resources for stu - dents during the learning process about medicine. Similarly, Hussain and Khan (2023) used data science to forecast aca - demic performance through machine learning algorithms. In particular, the linear regression and decision tree algorithms were used to analyse factors that influence academic performance in high school (Hussain & Khan, 2023). In ad - dition, Kostopoulos et al. (2021) used light gradient boosted, logistic regression and gradient boosting algorithms to predict grades during the use of massive open online courses (MOOCs). Data science allowed the identification of the relationship between stu - dent satisfaction and MOOCs through gradient boosting trees algorithm (Hew et al., 2020). In fact, this algorithm facilitated the construction of predictive models about MOOCs (Hew et al., 2020). Hsu-Wang (2019) analysed the behaviour of students in the Moodle platform considering the flipped classroom, that is, the linear regression algo - rithm allowed the prediction of school activities carried out in this technologi - cal tool both inside and outside the classroom. According to Nti et al. (2022), machine learning algorithms such as the decision tree and random forest al - lowed researchers to predict conditions concerning academic performance and social networks. Finally, educators and researchers use machine learning algorithms such as support vector machine, linear regression, k-nearest neighbors, decision tree, Bayesian classifier, random forest and deep learning to obtain new information about the relationship between educational phenomena and technology (Hus - sain & Khan, 2023; Lincke et al., 2021; Sghir et al., 2023). c e p s Journal 7 Research Problem At science faculties, teachers need to know students’ opinions on the use of technological tools in the educational field in order to organise and imple - ment new course activities. According to Sghir et al. (2023), the deep learn - ing algorithm allows the analysis of research hypotheses and the prediction of phenomena with great accuracy. Likewise, the random forest algorithm enables the construction of predictive models regarding technological and educational phenomena (Lincke et al., 2021; Sghir et al., 2023). The particular aims are: (1) to analyse the use of social networks and video conferencing systems for student autonomy and the exchange of ideas through the deep learning algorithm, (2) to build models of the use of these tools considering the random forest algorithm, and (3) to analyse students’ per - ceptions of social networks and video conferencing systems in the educational field. Research questions The aim of this mixed research is to analyse the perceptions of univer - sity students about the use of social networks and video conferencing systems through data science (deep learning and random forest algorithms). The re - search questions are: • How does the use of social networks and video conferencing systems in - fluence student autonomy and the exchange of ideas in the educational field (deep learning algorithm)? • What are the models on social networks and video conferencing systems to forecast student autonomy and the exchange of ideas in the educatio - nal field considering sex and age (random forest algorithm)? • What is the perception of university students about the use of social net - works and video conferencing systems in the educational field? The research hypotheses about the use of social networks and video con - ferencing systems are: • Hypothesis 1: Social networks positively impact student autonomy in the educational field. • Hypothesis 2: Social networks positively impact the exchange of ideas in the educational field. • Hypothesis 3: Video conferencing systems positively impact student au - tonomy in the educational field. analysis of social networks and video conferencing systems in the educational ... 8 • Hypothesis 4: Video conferencing systems positively impact the exchan - ge of ideas in the educational field. Method Participants The participants were 103 students of the Faculty of Sciences (sample 1) at the National Autonomous University of Mexico (NAUM) during the 2022 school year. In the study, sample 2 consisted of 21 teachers who were complet - ing a master’s degree in Upper Secondary Education at the NAUM during the 2025 school year. Instrument The study used the following model to analyse the use of social net - works and video conferencing systems through data science (See Figure 1). In the study, the independent variables are social networks and video conferenc - ing systems. The dependent variables are autonomy and the exchange of ideas, which are the target (predictive) variables used in the deep learning and ran - dom forest algorithms. Autonomy refers to student interaction in the teaching-learning process through social networks and video conferencing systems at any time, while exchange of ideas refers to communication and discussion of school topics through social networks and video conferencing systems from any location. Figure 1 Model on social networks and video conferencing systems The research used the deep learning algorithm to create forecasting models. Li et al. (2025) explain that this artificial intelligence technique uses neural network hyperparameters such as activation, hidden networks, cycles and sample size for the training and evaluation sections in order to establish c e p s Journal 9 predictive functions with great precision. Similarly, the random forest algo - rithm creates accurate predictive models, as it builds various trees until it finds the best option to forecast the phenomena (Beseiso, 2025; Jin, 2025). In the present study, the random forest algorithm allowed the construc - tion of the following models considering the characteristics of the participants: • Model 1 on social networks, profile of the students and autonomy. • Model 2 on social networks, profile of the students and the exchange of ideas. • Model 3 on video conferencing systems, profile of the students and autonomy. • Model 4 on video conferencing systems, profile of the students and the exchange of ideas. In December 2022, data collection was carried out at the National Au - tonomous University of Mexico through a questionnaire (See Table 1). This measurement instrument consists of four closed-ended questions about tech - nology, which were used in the quantitative approach. Two open-ended ques - tions about social networks and video conferencing systems were used in the qualitative approach. The closed questions were validated using the values ob- tained from the Load Factor > 0.600 and Composite Reliability > 0.700 (See Table 2). The scale of the responses was designed considering an article on data science published by Salas-Rueda et al. (2025a). analysis of social networks and video conferencing systems in the educational ... 10 Table 1 Measurement instrument about social networks and video conferencing systems No. Approach Scope Variable Dimension Question Answer 1 Quantitative Descriptive and causal Use of technology Social networks 1. Social networks facilitate learning Very much (1) Much (2) Little (3) Very little (4) Video conferencing systems 2. Video conferencing systems facilitate learning Very much (1) Much (2) Little (3) Very little (4) Autonomy 3. Digital tools facilitate student autonomy in the educational field Very much (1) Much (2) Little (3) Very little (4) Exchange of ideas 4. Digital tools facilitate the exchange of ideas in the educational field Very much (1) Much (2) Little (3) Very little (4) 2 Quantitative Descriptive Student Sex 5. Sex Man Woman Age 6. Age 20 years 21 years 22 years 23 years 24 years 25 years 3 Qualitative Students Use of social networks 7. What do you think of social networks? Open Use of video conferencing systems 8. What do you think of video conferencing systems? Open Teachers Social networks and video conferencing systems 9. What are the benefits of social networks and video conferencing systems in the educational field? Open Table 2 presents the validation of the measurement instrument on social networks and video conferencing systems. c e p s Journal 11 Table 2 Validation Variable Dimension Load Factor Average Variance Extracted Composite Reliability Use of technology Social networks 0.623 0.560 0.833 Video conferencing systems 0.665 Autonomy 0.806 Exchange of ideas 0.874 The research was supported by an Excel spreadsheet to calculate the frequencies of educational phenomena and the RapidMiner tool to calculate the deep learning and random forest algorithms. The deep learning algorithm requires the sample to be split in order to create the training and evaluation sections, while the random forest algorithm requires auxiliary variables such as sex and age in order to identify the relationship between the independent and dependent variables. Data science and machine learning algorithms allowed an analysis of the use of social networks and video conferencing systems during the COVID-19 post-pandemic through the RapidMiner tool (See Figure 2). In the deep learning algorithm, using the values of 60% ( n = 62), 70% (n = 72) and 80% ( n = 82) of the sample allows the construction of predictive models (training section), while using the values of 40% ( n = 41), 30% (n = 31) and 20% ( n = 21) of the sample identifies the most significant linear function to predict through the squared error (evaluation section). In the present study, hyperparameters on the size of the training and evaluation sections, cycles ( n = 10), hidden layers (50, 50) and Tanh activation were used to identify the best predictive models of autonomy and the exchange of ideas considering the use of social networks and video conferencing systems. Figure 2 RapidMiner tool analysis of social networks and video conferencing systems in the educational ... 12 Through the random forest algorithm, information about social net - works and video conferencing systems is used to build models considering the sex and age of the participants, while the objective variables are student au - tonomy and the exchange of ideas in the educational field. Finally, Excel allows descriptive analysis in the quantitative approach for questions about social networks, video conferencing systems, autonomy and exchange of ideas by calculating the frequencies. Research design The questionnaire was distributed to the students through Google Form under the direct supervision of the teacher, ensuring compliance with estab - lished ethical standards. The instructions were communicated in detail through Google Form. Participation in the study was entirely voluntary and no incen - tives were offered or provided. Results Digital tools facilitate student autonomy in the educational field very much ( n = 52, 50.49%), much ( n = 41, 39.81%) and little ( n = 10, 9.71%) (See Table 3). Digital tools facilitate the exchange of ideas in the educational field very much ( n = 38, 36.89%), much ( n = 43, 41.75%), little ( n = 18, 17.48%) and very little ( n = 4, 3.88%). Table 3 Results of social networks and video conferencing systems Question Answer n % 1. Social networks facilitate learning Very much (1) 34 33.01% Much (2) 51 49.51% Little (3) 16 15.53% Very little (4) 2 1.94% 2. Video conferencing systems facilitate learning Very much (1) 35 33.98% Much (2) 45 43.69% Little (3) 20 19.42% Very little (4) 3 2.91% 3. Digital tools facilitate student autonomy in the educational field Very much (1) 52 50.49% Much (2) 41 39.81% Little (3) 10 9.71% Very little (4) 0 0.00% c e p s Journal 13 Question Answer n % 4. Digital tools facilitate the exchange of ideas in the educational field Very much (1) 38 36.89% Much (2) 43 41.75% Little (3) 18 17.48% Very little (4) 4 3.88% 5. Sex Man 34 33.01% Woman 69 66.99% 6. Age 20 years 15 14.56% 21 years 4 3.88% 22 years 18 17.48% 23 years 46 44.66% 24 years 13 12.62% 25 years 7 6.80% The results of deep learning indicate that social networks and video conferencing systems positively impact student autonomy and the exchange of ideas (See Table 4). Table 4 Deep learning algorithm Hypothesis Training Deep learning Predictive model Conclusion Squared error (e 2 ) Value of p H1: Social networks → student autonomy 60% Cycles: 10 Hidden layers: 50, 50 Activation: Tanh y = 0.020x + 1.715 H1: Accepted 0.662 < 0.050 70% y = 0.064x + 1.592 H1: Accepted 0.716 < 0.050 80% y = 0.044x + 1.649 H1: Accepted 0.636 < 0.050 H2: Social networks → exchange of ideas 60% y = 0.031x + 1.811 H2: Accepted 0.845 < 0.050 70% y = 0.043x + 1.621 H2: Accepted 0.953 < 0.050 80% y = 0.131x + 1.901 H2: Accepted 0.819 < 0.050 H3: Video conferencing systems → student autonomy 60% y = 0.126x + 1.409 H3: Accepted 0.629 < 0.050 70% y = 0.055x + 1.404 H3: Accepted 0.673 < 0.050 80% y = 0.098x + 1.725 H3: Accepted 0.666 < 0.050 H4: Video conferencing systems → exchange of ideas 60% y = 0.373x + 1.120 H4: Accepted 0.758 < 0.050 70% y = 0.257x + 1.252 H4: Accepted 0.835 < 0.050 80% y = 0.380x + 1.098 H4: Accepted 0.720 < 0.050 Social networks Social networks facilitate learning very much ( n = 34, 33.01%), much ( n = 51, 49.51%), little ( n = 16, 15.53%) and very little ( n = 2, 1.94%) (See Table 3). The results of the deep learning algorithm with 60% (0.020), 70% (0.064) and analysis of social networks and video conferencing systems in the educational ... 14 80% (0.044) of the sample indicate that Hypothesis 1 is accepted (See Table 3). Social networks positively impact student autonomy. Figure 3 shows Model 1 on social networks to predict student autonomy. The random forest algorithm determined five predictive conditions. For ex - ample, if the student thinks that social networks facilitate learning very much and has an age > 20.5 years, then digital tools facilitate student autonomy in the educational field very much. Figure 3 Model 1 to predict student autonomy considering the use of social networks Age determines how social networks and the use of digital tools are re - lated for student autonomy. For example, if the student thinks that social net - works facilitate learning very much and has an age ≤ 20.5 years, then digital tools facilitate autonomy in the educational field much. The results of the deep learning algorithm with 60% (0.031), 70% (0.043) and 80% (0.131) of the sample indicate that Hypothesis 2 is accepted. Social networks positively impact the exchange of ideas. Figure 4 presents Model 2 on social networks to predict the exchange of ideas. The random forest algorithm determined seven predictive conditions. For example, if the student thinks that social networks facilitate learning very much and has an age > 20.5 years, then digital tools facilitate the exchange of ideas in the educational field very much. c e p s Journal 15 Figure 4 Model 2 to predict the exchange of ideas considering the use of social networks Sex and age determine how social networks and the use of digital tools for the exchange of ideas are related. For example, if the student thinks that social networks facilitate learning very much and has an age ≤ 20.5 years, then digital tools facilitate the exchange of ideas in the educational field much. According to the university students surveyed, social networks are use - ful to review the material, postings, comments and information of courses. “It is useful to keep informed about courses. ” “Social networks are useful for distance education. ” Social networks allow teachers and students to communicate from any - where during the learning process These technological tools also facilitate the dissemination of school information. “It serves to establish communication between teachers and students. ” “It is a good tool for the dissemination of information. ” In addition, social networks allow students to consult the multimedia resources of the courses. “Multimedia content is easier to handle. ” “When educational content is disseminated on the networks, the infor - mation is very attractive. ” Finally, the use of social networks promotes collaborative learning and the exchange of ideas in virtual modality. “They have become tools that allow collaborative learning, involve analysis of social networks and video conferencing systems in the educational ... 16 spaces for the exchange of information and encourage cooperation. ” “They are dynamic and serve to share interesting content. ” Video conferencing systems Video conferencing systems facilitate learning very much ( n = 35, 33.98%), much ( n = 45, 43.69%), little ( n = 20, 19.42%) and very little ( n = 3, 2.91%) (See Table 3). The results of the deep learning algorithm with 60% (0.126), 70% (0.055) and 80% (0.098) of the sample indicate that Hypothesis 3 is accepted (See Table 3). Video conferencing systems positively impact student autonomy. Figure 5 shows Model 3 on video conferencing systems to predict stu - dent autonomy. The random forest algorithm determined five predictive condi - tions. For example, if the student considers that video conferencing systems fa - cilitate learning very much and has an age > 21 years, then digital tools facilitate autonomy in the educational field very much. Figure 5 Model 3 to predict student autonomy considering the use of video conferencing systems Age determines how video conferencing systems and the use of digital tools are related for student autonomy. For example, if the student considers that video conferencing systems facilitate learning much and has an age > 20.5 years, then digital tools facilitate autonomy in the educational field very much. The results of the deep learning algorithm with 60% (0.373), 70% (0.257) and 80% (0.380) of the sample indicate that Hypothesis 4 is accepted (See Ta - ble 3). Therefore, video conferencing systems positively impact the exchange of ideas. c e p s Journal 17 Figure 6 shows Model 4 on video conferencing systems to predict the exchange of ideas. The random forest algorithm determined seven predictive conditions. For example, if the student considers that video conferencing sys - tems facilitate learning very much and has an age > 21 years, then digital tools facilitate the exchange of ideas in the educational field very much. Figure 6 Model 4 to predict the exchange of ideas considering the use of video conferencing systems Age and sex determine how video conferencing systems and the use of digital tools for the exchange of ideas are related. For example, if the student considers that video conferencing systems facilitate learning much and has an age ≤ 20.5 years, then digital tools facilitate the exchange of ideas in the educa - tional field much. According to the university students surveyed, video conferencing sys - tems facilitate communication and allow interaction. “I consider that it is the main means of communication. It is the closest thing to a class between the students and teacher. ” “Necessary applications for the development of distance education. ” As mentioned by the participants in this study, video conferencing sys - tems allow classes to be recorded. Zoom has an interface that is very simple to use during the educational process. “Zoom is great since it allows recording the session. The interaction is not complicated. It is very intuitive to use. ” “They are wonderful, it is very easy to create a meeting. ” analysis of social networks and video conferencing systems in the educational ... 18 Video conferencing systems allow a larger number of students to enrol in courses. Moreover, this technological tool allows the recording of classes, which can be consulted at any time. “They are extremely useful because they allow a greater number of peo - ple to take the courses. In addition, they allow the creation of support material for the following generations. ” “They are very useful for taking classes anywhere. ” According to the participants of the study, video conferencing systems facilitate the realisation of classes in a virtual modality. “Very efficient for online classes. ” “They are very useful for distance classes. ” Finally, the study presents teachers’ responses to the question: What are the benefits of social networks and video conferencing systems in the educa - tional field? According to the teachers surveyed, social networks and video confer - encing systems facilitate learning at any time. “These tools allow for more interactive work with students remotely. ” “They help students learn among peers. ” These technological tools also facilitate the creation of virtual environ - ments in which the student is the fundamental focus of the educational process. “These tools allow access to information at any time and encourage stu - dents to take an active role. ” “They help students act actively. ” The benefits of the use of social networks and video conferencing sys - tems include autonomy, motivation and participation during the teaching- learning process. “They foster autonomy in the user, spark interest and facilitate the re - trieval of information that requires feedback. ” “These applications motivate students during the learning process and encourage participation. ” “They improve autonomy and skill development among students. ” Teachers also believe that these communication tools facilitate the ac - cess and dissemination of content during the educational process. “The availability of materials online. ” c e p s Journal 19 “ Accessibility of classroom materials. ” “They are easy to use and available at any time of day. ” Finally, social networks and video conferencing systems facilitate com - munication between participants of the educational process from anywhere. “Improved the teacher-student communication. ” “To answer questions at different times. Communication is immediate between the participants. ” Discussion Educators and institutions are currently transforming activities through new technological developments (Al-Balushi et al., 2022; Cerda-González et al., 2022 ; Dominguez-Castillo et al., 2022 ; Lena-Acebo et al., 2023). In the pre - sent study, 90.30% of the participants think that digital tools facilitate student autonomy in the educational field very much or much. Likewise, 78.64% of the students surveyed consider that digital tools facilitate the exchange of ideas in the educational field very much or much. Most of the participating university students have a favourable opinion on these aspects. Social networks Today, teachers are looking for new tools such as social networks to create learning environments (Al-Balushi et al., 2022; Bendayan et al., 2024; García-García et al., 2023; Lundgren et al., 2022). In the present study, 82.52% of the students surveyed consider that social networks facilitate learning very much or much. In educational institutions, the incorporation of social networks has in - creased due to their availability and ease of use (Bendayan et al., 2024; Lena- Acebo et al., 2023; Lundgren et al., 2022). According to the university students surveyed, social networks are useful to review the material, postings, comments and information of the courses. The results regarding Hypothesis 1 indicate that social networks positively impact student autonomy in the educational field, with a value of p < 0.050. The participating teachers believe that the use of social networks and video conferencing systems allows the creation of virtual environments in which the student is the fundamental focus of the educational process. The students of the National Autonomous University of Mexico men - tion that social networks allow the communication and dissemination of analysis of social networks and video conferencing systems in the educational ... 20 information. In Model 1, the random forest algorithm determined six condi - tions considering the students’ profile. In particular, age determines how social networks and student autonomy are related. Furthermore, this model establish - es three predictive conditions where digital tools facilitate student autonomy in the educational field very much. As mentioned by Hasiloglu et al. (2020), teachers use social networks as tools to encourage interaction and collaboration. Similarly, the use of social networks at the National Autonomous University of Mexico allows students to consult the multimedia resources of the courses. The results on Hypothesis 2 indicate that social networks positively impact the exchange of ideas in the educational field, with a value of p < 0.050. In the present study, the educators surveyed state that the use of social networks and video conferencing systems in the teaching-learning process favours communication between participants from anywhere. According to the participants, the use of social networks promotes col - laborative learning and the exchange of ideas in virtual modality. In Model 2, the random forest algorithm determined seven conditions considering the stu - dents’ profile. In particular, age and sex determine how social networks and the exchange of ideas are related. Furthermore, this model establishes four predic - tive conditions where digital tools facilitate the exchange of ideas in the educa - tional field very much. Video conferencing systems As mentioned by Camilleri and Camilleri (2022), video conferencing systems such as Google Meet, Microsoft Teams and Zoom are technological tools that facilitate the presentation of school topics from anywhere. In the pre - sent study, 77.67% of the students surveyed think that video conferencing sys - tems facilitate learning very much or much. As Bailey (2022) establishes, the use of Zoom facilitates understanding of topics and analysis of school content. According to the students surveyed in the present study, video conferencing systems facilitate communication and in - teraction during virtual classes. The results on Hypothesis 3 indicate that video conferencing systems positively impact student autonomy in the educational field, with a value of p < 0.050. The teachers participating in the study believe that social networks and video conferencing systems facilitate the access and dissemination of school content from anywhere. According to the participants, one of the advantages of video conferencing systems is the recording of classes to review the topics. In Model 3, the random c e p s Journal 21 forest algorithm determined six conditions considering the students’ profile. In particular, age determines how video conferencing systems and student auton - omy are related. Furthermore, this model establishes two predictive conditions where digital tools facilitate autonomy in the educational field very much. Zoom is a communication tool that allows the resolution of doubts in real time (Faner et al., 2022; Unal & Yilmaz, 2024; Walcott-Bedeau, 2022). Moreover, video conferencing systems have a simple, fast and useful interface for the educational field. The results on Hypothesis 4 indicate that video con - ferencing systems positively impact the exchange of ideas in the educational field, with a value of p < 0.050. In this sense, the educators from the NAUM believe that social networks and video conferencing systems facilitate learning at any time. According to the participants of the present study, video conferencing systems facilitate the realisation of classes in a virtual modality. In Model 4, the random forest algorithm determined seven conditions considering the stu - dents’ profile. In particular, age and sex determine how video conferencing sys - tems and the exchange of ideas are related. Furthermore, this model establishes two predictive conditions where digital tools facilitate the exchange of ideas in the educational field very much. Finally, the training section was modified with values of 60%, 70% and 80% of the sample to identify the best function to predict autonomy and the exchange of ideas. The lowest value related to the squared error (e 2 ) allows establishing the most significant function to predict these events, that is, the function of y = 0.044x + 1.649 with e 2 = 0.636 for model 1, the function of y = 0.131x + 1.901 with e 2 = 0.819 for model 2, the function of y = 0.126x + 1.409 with e 2 = 0.629 for model 3, and the function of y = 0.380x + 1.098 with e 2 = 0.720 for model 4. Conclusion Universities are currently changing their teaching strategies supported by technological advances. In this study, the use of social networks and video conferencing systems favour the creation of educational spaces where the stu - dent is the protagonist of the teaching-learning process. The deep learning algorithm allows the analysis of technological and educational phenomena in order to construct predictive models considering hyperparameters such as activation, training and evaluation section sizes, cy - cles and hidden layers. Similarly, the random forest algorithm facilitates the construction of trees to predict events. analysis of social networks and video conferencing systems in the educational ... 22 The results of the deep learning algorithm indicate that social networks and video conferencing systems positively impact student autonomy and the exchange of ideas. This machine learning algorithm used the smallest value of the squared error to determine the forecast considering these technologi - cal tools. Educational interventions designed considering the support of these technological tools will therefore create a teaching-learning space where the student is autonomous and exchanges ideas. This research proposes various functions for predicting student auton - omy and the exchange of ideas based on the use of social networks and video conferencing systems in the educational context. All of the predictive models indicate a positive relationship between these study variables. The random forest algorithm allowed the analysis of these technological tools considering sex and age. This machine learning algorithm used student autonomy and the exchange of ideas as the target or prediction variables. The limitations of this mixed study are the dependent variables, the educational strategies and the sample size. Future research could analyse the use of social networks and video conferencing systems considering the active role, the development of skills, participation and collaborative work in various universities. Similarly, researchers could create predictive models that consider educational strategies through the deep learning and random forest algorithms. Predictive models related to the use of social networks and video confer - encing systems could consider the opinion of educators in order to predict vari - ables associated with the teaching-learning process. Similarly, trees obtained from the random forest algorithm could use the characteristics of teachers to find various relationships between technology and the educational process. Social networks facilitate use of multimedia resources, publication of school content and review of information, while video conferencing systems facilitate the realisation of classes in virtual modality through recordings and interaction. The present mixed research recommends the use of social net - works and video conferencing systems during the organisation of educational interventions, as these technological tools favour the creation of virtual envi - ronments for teaching and learning. Social media and video conferencing systems facilitate interaction be - tween participants of the educational process. The results of the deep learning and random forest algorithms indicate that educators can use these technologi - cal tools to plan creative and dynamic activities that foster student autonomy and the exchange of ideas. The surveyed teachers’ opinions on the use of social networks and video conferencing systems indicate that these tools facilitate learning at any time, c e p s Journal 23 access and dissemination of the content, creation of virtual environments and communication between participants in the educational process from anywhere. In conclusion, teachers can incorporate social networks and video con - ferencing systems into the educational field to organise and implement new virtual spaces where students become the axis of the learning process. Ethical statement This research study was approved by the Ethical Research Committee of the Institute of Applied Sciences and Technology, National Autonomous Uni - versity of Mexico. Disclosure statement The authors have no conflict of interest to declare. Acknowledgement This work thanks the Data Science and Artificial Intelligence Laboratory for its support. References Al-Balushi, W ., Al-Busaidi, F. S., Malik, A., & Al-Salti, Z. (2022). Social media use in higher educa - tion during the COVID-19 pandemic: A systematic literature review. International Journal of Emerg - ing Technologies in Learning, 17(24), 4–24. https://doi.org/10.3991/ijet.v17i24.32399 Ayanwale, M. A., Molefi, R. R. & Oyeniran, S. (2024). Analyzing the evolution of machine learning integration in educational research: A bibliometric perspective. Discover Education, 3, Article 47. https://doi.org/10.1007/s44217-024-00119-5 Bailey, D. (2022). Interactivity during Covid-19: Mediation of learner interactions on social presence and expected learning outcome within videoconference EFL courses. Journal of Computers in Educa - tion, 9, 291–313. https://doi.org/10.1007/s40692-021-00204-w Bendayan, M., Bonneau, C., & Delespierre, M. T., Sais, E., Picard, F., Alter, L., Boitrelle, F. & Cazabat, L. (2024). Evaluating the satisfaction and utility of social networks in medical practice and continu- Evaluating the satisfaction and utility of social networks in medical practice and continu - ing medical education. BMC Medical Education, 24, Article 186. https://doi.org/10.1186/s12909-024-05149-z Beseiso, M. (2025). Enhancing student success prediction: A comparative analysis of machine learn - ing technique. TechTrends, 69, 372–384. https://doi.org/10.1007/s11528-025-01044-6 analysis of social networks and video conferencing systems in the educational ... 24 Camilleri, M. A., & Camilleri, A. C. (2022). The acceptance of learning management systems and video conferencing technologies: Lessons learned from COVID-19. Technology, Knowledge and Learning, 27, 1311–1333. https://doi.org/10.1007/s10758-021-09561-y Caratiquit, K. D., & Caratiquit, L. J. C. (2023). Influence of social media addiction on academic achievement in distance learning: Intervening role of academic procrastination. Turkish Online Journal of Distance Education , 24(1), 1–19. Cerda-González, C., León-Herrera, M., Saiz-Vidallet, J. L., & Villegas-Medrano, L. (2022). Chilean student teachers’ purposes of use of digital technologies: Construction of a scale based on digital competences. Pixel-Bit, (64), 7–25. https://doi.org/10.12795/pixelbit.93212 Chen, F. (2022). Prediction, monitoring, and management of the classified training quality of english majors based on support vector machine. International Journal of Emerging Technologies in Learning , 17(24), 233–248. https://doi.org/10.3991/ijet.v17i24.35943 Cvitković, D., Wagner-Jakab, A., & Stošić, J. (2024). Remote learning and stress in mothers of students with attention deficit and hyperactivity disorder during the Covid-19 Lockdown. Center for Educational Policy Studies Journal , 14(3), 171–187. https://doi.org/10.26529/cepsj.1496 Dominguez-Castillo, J. G., Cisneros-Cohernour, E. J., Ortega-Maldonado, A., & Ortega-Carrillo, J. A. (2022). Students perceptions about distance learning during COVID-19. Pixel-Bit, (65), 237–273. https://doi.org/10.12795/pixelbit.94070 Faner, M. A., Ritchie, R. P ., Ruger, K. M., Waarala, K. L., & Wilkins, C. A. (2022). Student perfor - mance in medical biochemistry and genetics: Comparing campus-based versus zoom-based lecture delivery. BMC Medical Education, 22, 1–12. https://doi.org/10.1186/s12909-022-03873-y García-García, F. J., López-Francés, I., & Molla-Esparza, C. (2023). Social network analysis for peer inclusion in undergraduate online discussions. Pixel-Bit, (66), 7–29. Guillén-Gámez, F. D., Linde-Valenzuela, T., Ramos, M., & Mayorga-Fernandez, M. J. (2022). Iden - tifying predictors of digital competence of educators and their impact on online guidance. Research and Practice in Technology Enhanced Learning, 17, 1–8. https://doi.org/10.1186/s41039-022-00197-9 Hasiloglu, M. A., Calhan, H. S., & Ustaoglu, M. E. (2020). Determining the views of the secondary school science teachers about the use of social media in education. Journal of Science Education and Technolog y, 29, 346–354. https://doi.org/10.1007/s10956-020-09820-0 Hew, K. F., Hu, X., Qiao, C., & Tang, Y . (2020). What predicts student satisfaction with MOOCs: A gradient boosting trees supervised machine learning and sentiment analysis approach. Computers & Education, 145, Article 103724. https://doi.org/10.1016/j.compedu.2019.103724 Hsu-Wang, F. (2019). On prediction of online behaviors and achievement using self-regulated learn - ing awareness in flipped classrooms. International Journal of Information and Education Technology , 9(12), 874–879. Hussain, S., & Khan, M. Q. (2023). Student-performulator: Predicting students’ academic perfor - mance at secondary and intermediate level using machine learning. Annals of Data Science, 10, 637–655. https://doi.org/10.1007/s40745-021-00341-0 Isman, A., Sabuncuoglu-Inanc, A., & Akinci-Cotok, N. (2023). An analysis of personal factors af - c e p s Journal 25 fecting learning motivation: A research on the online education process during Covid-19 period in Turkiye. Turkish Online Journal of Distance Education , 24(1), 88–108. Iyamuremye, A., Niyonzima, F. N., & Mukiza, J. (2024). Utilization of artificial intelligence and ma - chine learning in chemistry education: a critical review. Discover Education, 3, 95. https://doi.org/10.1007/s44217-024-00197-5 Jin, X. (2025). Predicting academic success: Machine learning analysis of student, parental, and school efforts. Asia Pacific Education Review. https://doi.org/10.1007/s12564-023-09915-4 Kastiro, L. A., Qusef, A. D., & Alsalhi, N. R. (2022). E-Learning service quality during COVID-19 pandemic from postgraduate students’ perspective in Jordan. International Journal of Emerging Tech - nologies in Learning, 17(24), 219–232. https://doi.org/10.3991/ijet.v17i24.36007 Kostopoulos, G., Panagiotakopoulos, T., Kotsiantis, S., Pierrakeas, C., & Kameas, A. (2021). Interpre - table models for early prediction of certification in MOOCs: A case study on a MOOC for smart city professionals. IEEE Access, 9, 165881–165891. https://doi.org/10.1109/ACCESS.2021.3134787 Koyuncu, I., Kilic, A. F., & Orhan-Goksun, D. (2022). Classification of students’ achievement via machine learning by using system logs in learning management system. Turkish Online Journal of Distance Education, 23(3), 18–30. https://doi.org/10.17718/tojde.1137114 Kusumaningdyah, R., Devetak, I., Utomo, Y ., Effendy, E., Putri, D., & Habiddin, H. (2024). Teaching stereochemistry with multimedia and hands-on models: The relationship between students scientific reasoning skills and the effectiveness of model type. Center for Educational Policy Studies Journal , 14(1), 171–197. https://doi.org/10.26529/cepsj.1547 Lena-Acebo, F. J., Pérez-Escoda, A., García-Ruiz, R., & Fandos-Igado, M. (2023). Social media and smartphones as teaching resources: Spanish teacher’s perceptions. Pixel-Bit, (66), 239–270. Li, T., Haudek, K., & Krajcik, J. (2025). Utilizing deep learning ai to analyze scientific models: Over - coming challenges. Journal of Science Education and Technology , 34, 866–887. https://doi.org/10.1007/s10956-025-10217-0 Lincke, A., Jansen, M., Milrad, M., & Berge, E. (2021). The performance of some machine learning approaches and a rich context model in student answer prediction. Research and Practice in Technol- ogy Enhanced Learning, 16, Article 10. https://doi.org/10.1186/s41039-021-00159-7 Lundgren, L., Crippen, K. J., & Bex, R. T. (2022). Social media interaction as informal science learn - ing: A comparison of message design in two niches. Research in Science Education, 52, 1–20. https://doi.org/10.1007/s11165-019-09911-y Muls, J., De-Backer, F., Thomas, V ., Zhu, C., & Lombaerts, K. (2020). Facebook class groups of high school students: Their role in establishing social dynamics and learning experiences. Learning Envi- ronments Research, 23, 235–250. https://doi.org/10.1007/s10984-019-09298-7 Nguyen, X., Pho, D., Luong, D., & Cao, X. (2021). Vietnamese students’ acceptance of using video conferencing tools in distance learning in Covid-19 pandemic. Turkish Online Journal of Distance Education, 22(3), 139–162. https://doi.org/10.17718/tojde.961828 Nti, I. K., Akyeramfo-Sam, S., Bediako-Kyeremeh, B., & Agyemang, S. (2022). Prediction of social media effects on students’ academic performance using machine learning algorithms (MLAs). Jour - analysis of social networks and video conferencing systems in the educational ... 26 nal of Computers in Education, 9, 195–223. https://doi.org/10.1007/s40692-021-00201-z Quesnelle, K. M., & Montemayor, J. R. (2020). A multi-institutional study demonstrating under - graduate medical student engagement with question-type Facebook posts. Medical Science Educator, 30, 111–115. https://doi.org/10.1007/s40670-019-00910-2 Rice, M. F. (2022). Special education teachers’ use of technologies during the COVID-19 era (Spring 2020—Fall 2021). TechTrends, 66, 310–326. https://doi.org/10.1007/s11528-022-00700-5 Sahin-Dogruer, S. (2023). At school or home? Eight graders’ first practices with online geometry les - sons. Turkish Online Journal of Distance Education , 24(1), 220–233. Salas-Rueda, R. A., Domínguez-Herrera, E., & Castañeda-Martínez, R. (2025a). Padlet: A virtual wall to improve the teaching-learning process at the higher education level? Texto Livre, 18, Article e49123. https://doi.org/10.1590/1983-3652.2025.49123 Salas-Rueda, R. A., González-García, H., & Becerra-Torres, E. (2025b). Analysis of the educational application on mathematics for the Bachelor’s Degree in Applied Geography considering Data Sci - ence. Revista Sociedad & Tecnología, 8(2), 293–303. Sghir, N., Adadi, A., & Lahmer, M. (2023). Recent advances in predictive learning analytics: A decade systematic review (2012-2022). Education and Information Technologies, 28, 8299–8333. https://doi.org/10.1007/s10639-022-11536-0 Shiao, Y . T., Chen, C. H., & Wu, K. F. (2023). Reducing dropout rate through a deep learning model for sustainable education: Long-term tracking of learning outcomes of an undergraduate cohort from 2018 to 2021. Smart Learning Environments, 10, Article 55. https://doi.org/10.1186/s40561-023-00274-6 Unal, R., & Yilmaz, M. B. (2024). Accepting video conferencing technology as an in-service training tool for health professionals. Education and Information Technologies, 29, 21217–21239. https://doi.org/10.1007/s10639-024-12724-w Walcott-Bedeau, G. (2022). A pilot study to determine if playing music before class enhanced the “Zoom” online learning environment in a preclinical science course. Medical Science Educator, 32, 947–952. https://doi.org/10.1007/s40670-022-01596-9 Žabkar, J., Urankar, T., Javornik, K., & Košak-Babuder, M. (2023). Identifying reading fluency in pupils with and without dyslexia using a machine learning model on texts assessed with a readability application. Center for Educational Policy Studies Journal , 13(4), 233–256. https://doi.org/10.26529/cepsj.1367 c e p s Journal 27 Biographical note Ricardo-Adán Salas-Rueda, PhD, is a full-time academic at Na - tional School of Earth Sciences, National Autonomous University of Mexico, Mexico. Head of the Data Science and Artificial Intelligence Laboratory. His main areas of research are data science in education, pedagogical strategies sup - ported by technology, and intelligent educational applications. Erika-Patricia Salas-Rueda, PhD, studied at Monterrey Institute of Technology and Higher Studies, Mexico. Member of the Data Science and Artificial Intelligence Laboratory. Her main areas of research are data science in education, pedagogical strategies supported by technology, and intelligent educational applications. Rodrigo-David Salas-Rueda studied at Metropolitan Autonomous University, Mexico. Member of the Data Science and Artificial Intelligence Laboratory. His main areas of research are data science in education, pedagogi - cal strategies supported by technology, and intelligent educational applications.