Finding out topics in educational materials using their components

The Web is witnessing an exponential growth of distributed and heterogeneous educational material. This hampers distinguishing among contents of these materials, as well as their retrieval. While information retrieval and classification mechanisms concentrate on corpus analysis, annotation approaches either target specific formats or require that a document follows interoperable standards. Rather than target only textual characteristics, our strategy is mainly based on components of educational material. The header, body, footer and numbering of slides and progress bar are examples of components of slides and videos. Though our work is general purpose, it is being tested against slides and videos from Coursera, a web platform that provides universal access to online education material and courses from universities and organizations around the world.