Predicting Student’s Attributes from their Physiological Response to an Online Course

Marco Antonio Hernández Pérez, Emmanuel Rosado Martínez, Rolando Menchaca Méndez, Ricardo Menchaca Méndez, Mario E. Rivero Ángeles, Víctor M. González


In this work, we present the results of a study where we monitored the physiological response of a set of fifty high-school students during their participation in an online course. For each of the subjects, we recollected time-series obtained from sensors of physiological signals such as electrical cerebral activity, heart rate, galvanic skin response, body temperature, among others. From the first four moments (mean, variance, skewness and kurtosis) of the time series we trained Artificial Neural Network and Support Vector Machine models that showed to be effective for determining the sex of the subjects, as well as the type of activity they were performing, her learning style and whether or not they had previous knowledge about the course contents. These results show that the physiological signals contain relevant information about the characteristics of a user of an online learning platform and that this information can be extracted to develop better online learning tools.


Machine learning, electroencephalography, physiological response, e-learning

Full Text: PDF (Spanish)