Intelligent decision support systems for managing the diffusion of social computing in school-based ubiquitous learning

Abstract

The past decade has seen an explosion in social media applications. Most adolescents in South Africa have access to social media applications despite the country’s economic inequalities. The drive for social media applications is important to enhance human connectedness. In unprecedented times social computing can be utilised in school-based learning to benefit learners. Climate change has propagated extreme weather patterns which has increased the occurrence of natural disasters and diseases. The emergence of the novel Coronavirus resulted in most countries implementing nation-wide forms of lockdown to curb the spread of infection. Consequently, these adverse phenomena across the globe are disruptive to conventional schoolbased education. Ubiquitous learning (u-learning) relates to learning that occurs at any place without time constraints. In some schools, u-learning has become a conventional learning approach and pedagogy but there are various education and technology attributes that must be addressed for the penetration of social computing in schools. Therefore, there is a need to guide learners and school-based instructors on their preferences of digital access and social media applications. The main aim of the study was to investigate social media-driven Intelligent Decision Support Systems using live data, to assist instructors and learners manage the diffusion of social computing in school-based ubiquitous learning. In pursuing this study, a quantitative research methodology was used for the collection of data from learners and instructors from the schools in the eThekwini Region, namely, Umlazi District and Pinetown District of KwaZulu-Natal Province, South Africa. A survey was conducted to elicit data from participants on their use of social computing for u-learning. The approximate target population size was 129 421 individuals with a sample size of 384 participants. There were 260 respondents with an acceptable response rate of 67,71%. The study derived attributes for ranking the social media applications and Principal Component Analysis which is an unsupervised Machine Learning algorithm reduced the dimensionality of the attributes. The multi-criteria decision-making algorithm, Fuzzy Technique of Order Preference Similarity Ideal Solution was implemented to rank the social media applications in line with the dimensionality reduced criteria based on the subjective decisions of expert decision makers. Data Envelopment Analysis, another multi-criteria analysis method was utilised to score the efficiency of the devices for u-learning. The results showed the most precise, mathematically approved social media applications and devices that can support u-learning in schools. An automated application based on research evidence using Intelligent Decision Support Systems was designed as a research output.