Assessment of Modified BAT Algorithm for MOOC Learner Influence Maximization

Identification of a small group of individuals based on their maximal influence cascade is influence maximization. During the COVID-19 pandemic, discussion forums on the Massive Open Online Course (MOOC) platform have become a paramount interaction medium among learners, and the identification of influential learners evolved as a substantial research issue. In this research paper, an optimization function based on an independent cascade is established for the discussion forum influence maximization problem. A modified version of the BAT algorithm is proposed which memorizes the bad experience of the BAT. The proposed Modified algorithm helps the BAT to remember the worst location that has already been traversed for a reliable estimation in an optimized manner while exploring the best solution. Further, the performance of BAT and Modified BAT for influence maximization on the discussion forum network of a MOOC platform is evaluated which shows the excellent performance of modified BAT. Convergence graph for different populations on deviating probability depicts the effective performance of modified BAT over generic BAT algorithm.