Özet:

Abstract The increasing use of e-learning by students causes an LMS to consider student learning styles to provide comfortable content and improve the learning process. Learning style refers to the preferred way in which an individual learns in the best way. The traditional method for detecting learning styles (using questionnaires) has many limitations, namely the process of filling out the questionnaire is time consuming, and the results obtained are inaccurate because students are not always aware of their own learning preferences. So, in this study we use an approach to detect learning styles automatically, based on the Felder and Silverman learning style model (FSLSM) and use a machine learning algorithm. The proposed approach consists of two parts: The first part aims to extract the sequence of student activities from the log file, map with literature based then use an unsupervised algorithm (K-means) to group them into sixteen clusters according to FSLSM, and the second part uses a supervised algorithm (Naive Bayes) to predict learning styles for new activity sequences or new students. To take this approach, we use real datasets extracted from e-learning system log files. To evaluate performance, we used the confusion matrix. The more learning activities will increase the features and increase accuracy.

Anahtar Kelimeler: