Exploring the Navigation Patterns of Learners on an Educational Recommender System

In this study, the interaction of students with an educational recommender system has been analyzed based on log data. Considering the fact that it is quite a challenging task to measure and model how learning actually takes place, the main contribution of this paper is to provide insights into the learning processes of learners based on interaction data. Various types of factors about learners such as prior knowledge level, motivation level, and learning styles can affect their learning processes and the learning outcomes directly. These factors also affect the usability of educational systems among other system-related aspects. In this regard, analyzing the usage data of adaptive e-learning systems can provide valuable information about the effectiveness of recommendations and the learning processes of learners. With this motivation, we collected and analyzed the interaction data of a semantic web-based educational recommender system. The educational recommender system was used as a supplementary tool by 172 voluntary students who took the Digital Computer Design course of the Computer Engineering curriculum at a public university in Turkey. Different types of learning resources such as presentations, videos (for teaching theoretical subjects and demonstrating different examples), exercises, and quizzes have been used as learning materials. Users can directly access the learning materials in whichever order they prefer or may follow the recommendations of the e-learning system. We analyzed the system interaction data to provide descriptive statistics and insights on various aspects such as learning material usage patterns of learners, popularity of learning materials based on learning material types, success levels at quizzes, and whether the learners were likely to follow recommendations given as hints after wrong answers at quizzes. The results are promising; however, considering the limited application size of a single course, further studies should be conducted to support the findings.