Characterization of learning styles as student modeling to determine the zone of proximal development for content selection in an intelligent tutorial system.
Student modeling in an intelligent tutorial system has been approached through knowledge tracking; however, one challenge is integrating pedagogical models into the construction of intelligent tutorials. Machine learning allows students to be modeled using historical data to predict learning outcomes. Neural networks within machine learning perform classification through a training process. This work characterizes student modeling with historical data on learning styles to model students and determine their Zone of Proximal Development (ZPD). The ZPD is determined using a Kohonen neural network model, which generates a self-organizing map (SOM) that allows us to classify a type of student modeled by their learning style. The Kohonen network was trained with a 6 x 6 grid, giving a total of 36 neurons and 803 vectors, of which 70% = 562 were used for training and 30% = 241 for testing. The self-organizing map generated four classes of ZPD into which the learning style can fall.
Characterization of learning styles as student modeling to determine the zone of proximal development for content selection in an intelligent tutorial system.
-
DOI: https://doi.org/10.22533/at.ed.8208112620017
-
Palavras-chave: : Intelligent tutoring systems, Student modeling, Learning styles, Neural networks, Zone of Proximal Development
-
Keywords: : Intelligent tutoring systems, Student modeling, Learning styles, Neural networks, Zone of Proximal Development
-
Abstract:
Student modeling in an intelligent tutorial system has been approached through knowledge tracking; however, one challenge is integrating pedagogical models into the construction of intelligent tutorials. Machine learning allows students to be modeled using historical data to predict learning outcomes. Neural networks within machine learning perform classification through a training process. This work characterizes student modeling with historical data on learning styles to model students and determine their Zone of Proximal Development (ZPD). The ZPD is determined using a Kohonen neural network model, which generates a self-organizing map (SOM) that allows us to classify a type of student modeled by their learning style. The Kohonen network was trained with a 6 x 6 grid, giving a total of 36 neurons and 803 vectors, of which 70% = 562 were used for training and 30% = 241 for testing. The self-organizing map generated four classes of ZPD into which the learning style can fall.
- José Antonio Cruz Zamora
- Blanca Estela Pedroza Méndez
- Yesenia Nohemí González Meneses
- Rodolfo Eleazar Pérez Loaiza
