Adaptive Quizzes Using Fuzzy Genetic Algorithm
Spyros Papadimitriou, Konstantina Chrysafiadi, Maria K. Virvou · 2023
The quizzes allow the tutor to assess the student's knowledge level briefly. However, creating a quiz is a challenging task that requires time and effort. If the tutor wants to create multiple quizzes tailored to the student's needs, the time needed to construct them increases dramatically. Furthermore, a large pool of questions and activities makes creating personalised quizzes more complex. A solution to the above is employing genetic algorithms to quickly and automatically produce personalised quizzes and tests. However, a genetic algorithm uses complex mathematical forms, which tutors of various knowledge domains find difficult to understand. Given this, the paper presents a genetic algorithm that embeds a fuzzy rule-based mechanism in the fitness calculation phase to select the most appropriate questions for each learner and create personalised quizzes more realistically. The presented algorithm aims to describe with linguistic terms the appropriateness of each question rather than obscure mathematics forms. In such a way, estimating the suitability of a quiz question is closer to the human way of thinking and becomes more explainable to tutors. The presented fuzzy genetic algorithm has been performed for creating quizzes for learners of the programming language HTML. The algorithm accepts a pool of questions and the learner's characteristics and needs. It gives as output a report, which includes the questions sorted by their suitability and a recommended quiz that best fits the learner's needs.