oa Using information extraction to dynamically create multimedia tutorials

Background and Objectives: Multimedia tutorials are often useful for children with disabilities as they are better able to understand new concepts with the use of lessons that engage their senses. These tutorials should include images, sounds and videos. We propose a system to dynamically generate multimedia tutorials that can easily be customized by the instructor, with the use of domain specific information extraction. Methods: Text processing is performed with a stemming algorithm, after which formal concepts analysis is used to extract pre-specified keywords. A formal concept is represented as a hierarchical lattice structure and in this study is applied to the animal kingdom domain. Ontology-based information extraction is then performed, where multimedia elements are extracted online and mapped, by querying the Google image database. Results: The proposed system allows for automated, speedy and efficient dynamic generation of multimedia customized tutorials. Conclusions: This system automates the tutorial generation process and gives disabled children the opportunity to learn with tutorials designed to suit their intellectual needs.