Recommendation Model for Learning Material Using the Felder Silverman Learning Style Approach

Abstract

The biggest obstacle that students have when participating in a virtual learning environment (e-learning) is discovering a platform that has functionalities that can be customized to fit their needs. This is usually accomplished in several ways using educational resources such as learning materials and virtual classroom design elements. Our research has tried to meet this demand by suggesting an extra element in the virtual classroom design, i.e., classifying the students’ learning styles through machine-learning techniques based on information gathered from questionnaires. This feature allows teachers or instructors to modify their lesson plans to better suit the learning preferences of their students. Additionally, this feature aids in the creation of a learning path that serves as a guide for students as they choose their course materials. In this study, we have selected the Felder-Silverman Learning Style Model (FSLSM) in the questionnaire design, which focuses on identifying the students' learning styles. After that, we employ several machine learning algorithms to create a prediction model for the students’ learning styles. The algorithms include Decision Tree, Support Vector Machines, K-Nearest Neighbors, Naïve Bayes, Linear Discriminant Analysis, Random Forest, and Logistic Regression. The best prediction model from this exercise contributes to the recommendation model that was created using a collaborative filtering algorithm. We have carried out a pre-test and post-test method to evaluate our suggestions. There were 138 learners who were following a learning path and participated in this study. The findings of the pretest and post-test indicated a notable increase in students' motivation to study. This is confirmed by the fact that learners' satisfaction with online learning climbed to 87% when the learning style was considered, from 60% when it wasn't. Doi: 10.28991/HIJ-2023-04-04-010 Full Text: PDF