This paper presents the preliminary results of using machine learning techniques to analyze educational robotics activities. An experiment was conducted with 197 secondary school students in Italy: the authors updated Lego Mindstorms EV3 programming blocks to record log files with coding sequences students had designed in teams. The activities were part of a preliminary robotics exercise. We used four machine learning techniques—logistic regression, support-vector machine (SVM), K-nearest neighbors and random forests—to predict the students’ performance, comparing a supervised approach (using twelve indicators extracted from the log files as input for the algorithms) and a mixed approach (applying a k-means algorithm to calculate the machine learning features). The results showed that the mixed approach with SVM outperformed the other techniques, and that three predominant learning styles emerged from the data mining analysis.
Analysis of Educational Robotics Activities Using a Machine Learning Approach
Lect. Notes in Networks, Syst.
Makers at School, Educational Robotics and Innovative Learning Environments ; Chapter : 27 ; 203-211
2021-12-11
9 pages
Article/Chapter (Book)
Electronic Resource
English
Teaching Physics Concepts Using Educational Robotics
TIBKAT | 2020
|The Effectiveness of Educational Robotics Simulations in Enhancing Student Learning
Springer Verlag | 2023
|Revisiting the Pedagogy of Educational Robotics
Springer Verlag | 2023
|