A model to characterize exercises using probabilistic methods
Alberto Jiménez, Pedro J. Muñoz‐Merino, Carlos Delgado Kloos · Ninth International Conference on Technological Ecosystems for Enhancing Multiculturality (TEEM'21) · 2021
Many studies have been conducted on modeling learners in education using probabilistic methods to infer different indicators. However, little research has been done on modeling content in education by estimating different content characteristics using probabilistic methods. Based on the existing gap in this area, this paper presents a model for exercises using probabilistic methods with interactions carried out by students to obtain content characteristics and their relationship such as the number of attempts, grade, and time spent related with student efficiency. Simulations were performed to train the model and find the different curves that best characterize the exercises based on programmatically generated interactions. This model allows the teacher to redesign the exercises to improve the student’s learning process.