The group presented their paper entitled “Towards the modelling of the concentrated state of learners. An intra-subject approach”. The European Simulation and Modelling Conference (ESM) is one of the oldest European international conferences on computer simulation and modelling concerned with state of the art technology in modelling and simulation. Due to COVID restrictions in France the participation in the conference was online.
Proceedings can be found here.
In this work, we present an experimental study that aims to explore the potential application of Hidden Markov Models (HMM) to predict the state of concentration of learners. The features used for the prediction are extracted from 4 commonly used physiological signals. Previous works that pointed how emotion translates into different physiological responses from one subject to another have motivated the study of intra-subject learner models. Thus in this work both inter-subjects and intra-subjects approaches have been evaluated. To this end, we have built a labelled dataset from an intensive data capture on 2 different subjects that developed different learning tasks across several sessions. Results are consistent across
users and show that a high accuracy can be achieved when using HMM-based intra-subject models, but inter-subject models fail at the same task.