Article Preview
Top1. Introduction
With the extension of online education due to the COVID-19 situation, a plethora of massive open online course (MOOC) portals such as Coursera, Udacity, Udemy, Simplilearn, edX, Canvas Network, and Swayam are constantly drawing researchers’ attention (Seale et al., 2020). These online learning environments offer large-scale access to courses of all educational areas which helps students to reduce the cost of their professional education. The diverse nature of pedagogical materials is arranged in these MOOC courses, for example- Video lectures, Online quizzes, Lecture slides, Programming assignments, Online programming compilers. Undoubtedly, these MOOC portals have empowered individuals to gain knowledge from someone sitting far away and flourished to disseminate knowledge globally regardless of geographical and temporal boundaries. COVID situation heralded MOOC portals as a major attention-seeking innovation that drastically changed the future education system which was stagnant over the past so many decades.
This innovation gave birth to numerous research directions under this wide online learning education spectrum. Model-driven approaches are explored by researchers for the Technology Acceptance Model (TAM) (Fianu et al., 2020; Kurniawan et al., 2021), Student usage model for acceptance and use of technology (Fianu et al., 2020), etc. Further, in recent literature analysis-based knowledge-gathering approaches are implemented for course design factors analysis (Kim et al., 2021), evaluation of student feedback (Lundqvist et al., 2020), MOOC discussion Forum based feedback analysis (Onan & Toçoğlu, 2021) where natural language processing, learning techniques, and information retrieval techniques are examined. A promising and contemporary research area is Influence spreader identification which is performed for learners’ context influence on their MOOC learning (Hood et al., 2015). Out of various research directions, the underexplored and challenging area in the domain is the influence maximization of entities of the MOOC portal. This same area has been explored in this research work. MOOC influence maximization study can examine the influential course i.e., the course having a maximum influence on MOOC course learners, and can work towards identification of influential learners as well. This work is useful for viral marketing in diverse directions (Bana & Arora, 2018), some are - Influential Factor driven course promotion; Influential learner identification based on the number of courses viewed, explored, and certified by the learner; Course recommendation (Zhang et al., n.d.); Course - Course mapping network; suggestion for courses basket based on folded bipartite graph for course-learner; Same skill set learner identification based on formed folded bipartite graph of learner-learner. These research gaps exist due to the lack of data availability those needs to be filled out.
Like other social networks, the MOOC course-learner graph, Learner-Learner conversation graph are arbitrary in size and contain a complex topological structure. So classical social network representation characteristics solely are not sufficient to handle the influence maximization problem. One solution to the problem is to adopt the greedy approach to generate an influential course/learner seed set (Goyal et al., 2011a; WuGuanhao et al., 2021), but the system may cohere to locally optimal solutions that are not effective and global best solution. Henceforth, to get effective global optimal solution, Swarm Intelligence Algorithms- Particle swarm optimization (Gong et al., 2016), BAT (Tang et al., 2018a), Grey Wolf Algorithm (Zareie et al., 2020a), Artificial Bee Colony (A. Arora et al., 2019), Ant Colony Optimization (Singh et al., 2019); Evolutionary Algorithms- Differential Evolution (L. Qiu et al., 2021), Genetic Algorithm (Bucur & Iacca, 2016) have already been validated for numerous real-life influence maximization applications.