PREDICTING BEHAVIOUR CHANGE IN STUDENTS WITH SPECIAL EDUCATION NEEDS USING MULTI MODAL LEARNING ANALAYTICS

Abstract

The availability of educational data in novel ways and formats brings new 
opportunities to students with special education needs (SEN), whose behavior and 
learning are highly sensitive to their body conditions and surrounding environments. 
Multimodal learning analytics (MMLA) captures learner and learning environment 
data in various modalities and analyses them to explain the underlying educational 
insights. In this work, we applied MMLA to predict SEN students’ behavior change 
upon their participation in applied behavior analysis (ABA) therapies, where ABA 
therapy is an intervention in special education that aims at treating behavioral 
problems and fostering positive behavior changes. Here we show that by inputting 
multimodal educational data, our machine learning models and deep neural network 
can predict SEN students’ behavior change with optimum performance of 98% 
accuracy and 97% precision. We also demonstrate how environmental, psychological, 
and motion sensor data can significantly improve the statistical performance of 
predictive models with only traditional educational data. Our work has been applied 
to the Integrated Intelligent Intervention Learning (3I Learning) System, enhancing 
intensive ABA therapies for over 500 SEN students in Hong Kong and Singapore 
since 2020.