Preparing for Future Crises in Education: AI and Regression-Based Modeling of Emergency Distance Learning Satisfaction among Moroccan Law Students

Abstract

Higher education institutions around the world have been forced to use Emergency Distance Learning (EDL) in place of traditional classroom instruction due to the COVID-19 pandemic. Moroccan law students were impacted by this shift, experiencing issues with access, the caliber of their education, and their technical readiness. Using both traditional statistical techniques and advanced artificial intelligence (AI) methods, this study aims to identify and forecast the primary factors influencing students' satisfaction with EDL in Moroccan legal education. Using a thorough survey of 16,187 law students from public and private Moroccan universities, we investigate the effects of six predictor categories on overall satisfaction: technological readiness, accessibility, teacher quality, assessment methods, learning engagement, and awareness. We combine artificial intelligence models like Support Vector Regression (SVR), Random Forest (RF), LightGBM, and Multilayer Perceptron Regression (MLPR) with multiple linear regression and stepwise regression. To improve model correctness, features are chosen using Recursive Feature Elimination (RFE). Students expressed concerns about the fairness and transparency of online assessments and expressed a preference for in-person instruction despite significant technical readiness. This study highlights the importance of digital equality, pedagogical support, and intelligent analytics for strong higher education systems and provides evidence-based insights to improve the architecture of distance legal education in Morocco.