An Adaptive Quiz Generation System for Moodle using moosh and LLM

Abstract

This paper describes a system for automating adaptive quiz creation within Moodle. Leveraging `moosh` and the OpenAI API, it provides personalized learning. An initialization script sets up the course, while a periodically run script analyzes learner performance on existing quizzes. This analysis drives the generation of new, tailored quizzes that focus on individual weaknesses, offering remedial or supplementary exercises to enhance the learning experience. Helper functions are modularized for reusability.