With the ever-growing knowledge on the internet, students rely more on online materials to complement their course work. However, learners can easily be overwhelmed by the broad range of available materials. Concept maps have been proposed as a visualization framework to paint a broader picture about a topic by showing how different concepts are interconnected and related to the overall topic. Using this means of visualization has been shown to be beneficial for learners on multiple levels, such as improved knowledge retention, a more satisfactory learning experience, and better learning achievements. Prerequisite graphs are a special form of a concept map, where nodes represent concepts and an edge from concept c_i to concept c_j indicates that mastering the more basic concept c_i is needed before one may proceed with learning about the more advanced concept c_j. In other words, edges encode prerequisite relations. Having such information available when beginning to learn about a topic benefits learners in multiple ways. First, it acts as a guide to make learning a more satisfying experience, because it enables learners to continue with any preferred concept for which they already know all prerequisites. This way less frustration occurs while learning since learners do not face unknown or surprising prerequisite while learning about a specific concept as all prerequisites have been studied beforehand. Having potentially multiple concepts to choose from also allows individual learning paths through the material. Second, prerequisite graphs emphasize how seemingly unrelated concepts about a topic are connected with each other, which is one of the key takeaways from studying a topic.
Constructing prerequisite graphs is typically performed manually by experts, which is laborious and time-consuming. The main motivation for not automating the process of constructing prerequisite graphs from a given set of concepts with their textual descriptions is the seemingly inherent complexity of the task.
The aim of this project is to develop a methodology for constructing prerequisite relations among concepts based on learning materials in Turkish. Students will base their work on the methodology that was already developed at Sabancı University by the advisors. The current methodology is developed for the English Language and with this project, students will adopt it to Turkish by training the language model for Turkish. They will also integrate their work into the exisiting software tool, test it, and use it on a large scale set of educational resources in Turkish. The methdology that will be implemented in collaboration with MORPA (www.morpakampus.com) a company which provides a learning management system and educational materials for students in Turkey. MORPA will provide educational materials as well as experts in education to construct a prerequisite graph on a set of selected topics.
Students who will carry out the project should be familiar with Python and ideally with Web programming