Abstract:Education is a dynamic field that must be adaptable to sudden changes and disruptions caused by events like pandemics, war, and natural disasters related to climate change. When these events occur, traditional classrooms with traditional or blended delivery can shift to fully online learning, which requires an efficient learning environment that meets students' needs. While learning management systems support teachers' productivity and creativity, they typically provide the same content to all learners in a course, ignoring their unique learning styles. To address this issue, we propose a semi-supervised machine learning approach that detects students' learning styles using a data mining technique. We use the commonly used Felder Silverman learning style model and demonstrate that our semi-supervised method can produce reliable classification models with few labeled data. We evaluate our approach on two different courses and achieve an accuracy of 88.83% and 77.35%, respectively. Our work shows that educational data mining and semi-supervised machine learning techniques can identify different learning styles and create a personalized learning environment.

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is that the existing Learning Management Systems (LMS) fail to provide personalized learning content according to students' different learning styles. Specifically, although the traditional questionnaire - based method is considered reliable and effective, it has some limitations. For example, students' random answers may lead to inaccurate results, and the questionnaire lacks flexibility and is difficult to adapt to the changes in students' learning styles. To solve these problems, the paper proposes a semi - supervised self - taught labeling machine - learning method to detect students' different learning styles and create a personalized learning environment. This method utilizes the Felder - Silverman Learning Style Model (FSLSM) and has been proven through experiments that it can generate reliable classification models even with a small amount of labeled data. The following are the main contributions of this paper: 1. **Evaluating student behavior**: Reliably label and infer students' learning styles from the collected Moodle data. 2. **Semi - supervised self - training mechanism**: Use a small amount of labeled data to generate reliable classification models. 3. **Performance comparison**: Use different classifiers to compare the predictive ability of semi - supervised training models and evaluate their performance on independent data. 4. **Experimental verification**: Conduct experiments using real - data from two different courses and benchmark against similar works. Through these methods, the paper aims to bridge the gap between personalized learning and instructional design strategies, improve the quality of educational services, and meet the needs of different students. ### Key formulas and concepts - **Felder - Silverman Learning Style Model (FSLSM)**: - Four dimensions: Sensing - Intuitive (S - I), Visual - Verbal (V - V), Active - Reflective (A - R), Sequential - Global (S - G). - **Semi - supervised learning**: - Self - training: Train the model with the labeled data, then use the model to predict the labels of the unlabeled data, and then add the new labels to the training set for further training. - Co - training: Use multiple classifiers, and each classifier is trained based on different views of the data. - Graph - based methods: Construct a graph based on data similarity and propagate labels through the graph. - Entropy minimization: Reduce the uncertainty in the label prediction of unlabeled data. Through these techniques, the paper shows how to use machine learning and educational data mining to identify students' learning styles and support the creation of personalized learning environments.

Learning Style Identification Using Semi-Supervised Self-Taught Labeling

A fuzzy classification approach for learning style prediction based on web mining technique in e-learning environments

Predicting students' learning styles using regression techniques

Unsupervised Learning Style Classification for Learning Path Generation in Online Education Platforms

Machine Learning Approach for an Adaptive E-Learning System Based on Kolb Learning Styles

Predict student learning styles and suitable assessment methods using click stream

Semi-Supervised Feature Selection of Educational Data Mining for Student Performance Analysis

A learning style classification approach based on deep belief network for large-scale online education

A Hybrid Teaching Mode Based on Machine Learning Algorithm

FSLSM-Based Analysis of Student Performance Information in a Blended Learning Course Using Moodle LMS

The Discrimination of Learning Styles by Bayes-Based Statistics: an Extended Study on ILS System.

Learning Style Identification by CHAEA Junior Questionnaire and Artificial Neural Network Method: A Case Study

The Adaptive Learning System Based on Learning Style and Cognitive State

E-learning and COVID-19: Predicting Student Academic Performance Using Data Mining Algorithms

Identifying learning styles and cognitive traits in a learning management system

Students' performance in interactive environments: an intelligent model

Enhancing College Student Education and Management through Semisupervised Learning

Teacher/Student Deep Semi-Supervised Learning for Training with Noisy Labels

Exploring the Effectiveness of Learning Path Recommendation based on Felder-Silverman Learning Style Model: A Learning Analytics Intervention Approach

Discriminable Multi-Label Attribute Selection for Pre-Course Student Performance Prediction

Advancing Personalized and Adaptive Learning Experience in Education with Artificial Intelligence