Pixel International Conferences

This paper describes the design of an adaptive learning environment using statistics as the experimental domain. Providing better ways to learn statistics in higher education would improve statistics knowledge among graduates and make offering statistics courses cheaper. Statistics learning can demonstrate benefits of adaptation to (1) the subject domain in which the learner will apply statistics and (2) the mathematical knowledge of the learner. The system will create  individualized learning paths optimally adapted to the learner's goals. existing knowledge, cognitive abilities, learning style, and circumstances by recommending the next step throughout a learning interaction - look at another example suitable for this learner, solve another practice problem, or move on to the next concept. These recommendations draw on an extensive knowledge base using a combination of artificial intelligence and machine learning, of knowledge-based reasoning, and of learning analytics. The knowledge base includes statistical concepts and their relationships; learning objectives; many minimal presentation chunks - concept explanations, examples, problems to solve, questions - that can be sequenced in a learning path and that are indexed by concepts covered, learning objectives, prerequisites, difficulty, etc.; learners with rich profiles, including learner characteristics, history of progress through the system (including performance on tests after completing a learning path), and feedback on what presentations they liked. The learning analytics approach uses this data to derive even better learner profiles and conclusions about what learning materials suit a learner (and learners with similar profiles) to make more accurate predictions of the best next step in a learning path. The knowledge base can also be used to power a statistical advisor to recommend statistical methods and things the user should learn to apply these methods properly. Our hypothesis is that system following our design produces better learning outcomes; we plan to test this hypothesis.

Keywords: Individualized learning, Learning path recommendations, knowledge base of learning units, Learner characteristics, Learning analytics, Artificial intelligence;

References:

 
[1] Soergel, D. (2018). Innovative education enabled by Knowledge Organization and IT: Goal-directed, flexible, individualized, collaborative.
FOE 8, Venice, 2018-06-28/29. Proceedings . libreriauniversitaria. it Edizioni. 2018. 688 p. ISBN: 8833590208.  p. 273-278
[2] Soergel, D.; Ituralde Escudero, .F. (2018). Toward a Universal Document Model for Active Knowledge. ICKM 2018. https://digital.library.unt.edu/ark:/67531/metadc1393843/?q=soergel
[3] Wenger, E. (2014). Artificial intelligence and tutoring systems: computational and cognitive approaches to the communication of knowledge. Morgan Kaufmann.
[4] Aleven, V., et al. (2016). Help helps, but only so much: Research on help seeking with intelligent tutoring systems. International Journal of Artificial Intelligence in Education, 26(1), 205-223.