Abstract
One of the most vexing aspects of tertiary education is the learning behaviour of many beginners: Late drop-outs after much time has already been invested in attending a course, incomplete homework even though completed homework is a sufficient condition for success at examinations, and misconceptions that are not overcome early enough. This article presents three predictors related to these learning-impairing behaviours that have been built from data collected with a learning platform and by examining homework assignments, and developed as Hidden Markov Model, by relying on Collaborative Filtering, and by using Multiple Linear Regression. The sensitivities and specificities of the first two predictors are above 70% and the R-2-error of the third predictor is about 20%. Considering the large numbers of unknown parameters like course-independent learning, this quality is satisfying. The predictors have been developed for fostering a better learning by raising the learners' consciousness of the deficiencies of their learning. In other words, the predictors aim at "getting it wrong". The article reports on the predictors and their evaluation.
Item Type: | Journal article |
---|---|
Faculties: | Mathematics, Computer Science and Statistics > Computer Science |
Subjects: | 000 Computer science, information and general works > 004 Data processing computer science |
ISSN: | 2194-5357 |
Language: | English |
Item ID: | 82289 |
Date Deposited: | 15. Dec 2021, 15:01 |
Last Modified: | 15. Dec 2021, 15:01 |