A Presentation Mining Framework: From Text Mining to to Mind Mapping

Slide-based presentations, such as those created using PowerPoint, have become ubiquitous in educational settings, offering educators a convenient tool for conveying information to students. However, a significant challenge arises from the inherent linear structure of these presentations, which may not align with the non-linear process of knowledge reconstruction undertaken by students. This discrepancy often leads to variations in students’ understanding and interpretation of the presented material, potentially resulting in misinterpretation of the original content sequence of the slides. Recognizing this challenge, this paper proposes a presentation mining framework to address the gap between the linear knowledge construction by presenters and the non-linear knowledge reconstruction by students. By leveraging text mining techniques, the framework aims to uncover keywords and key phrases embedded within slide content, facilitating the generation of a visual representation in the form of mind map that captures the underlying structure and relationships of the extracted information. The mind maps generated are then evaluated against a domain ontology, a dictionary, and subject matter experts based on averaged precision, recall, and F-measure score. The results showed that the proposed framework is able to extract and generate reasonable visualization from presentation slides.