A Semi-automatic System to Detect Relevant Learning Content for Each Subject
Today, the crisis has worsened the panorama for Universities, placing new constraints that require being more sustainable economically. In addition, universities will also have to improve their research and teaching in order to obtain more research funds and attract more students. In this panorama, analytics can be a very useful tool since it allows academics (and university managers) to get a more thorough view of their context, to better understand the environment, and to identify potential improvements. Some analytics have been done under the names of Learning analytics, Academic analytics, Educational Data Mining and etcetera. However, these systems, under our humble opinion, only take into account the small part of data related to the problem, but not contextual data. In order to perform analytics efficiently and reproduce their results easily in other contexts, it is necessary to have as much information as possible about the context. For example, when communication forums are analyzed to see the concepts in which students have more doubts, it is important to analyze also what concepts are taught in the subject. Having access to both sources, more information can be obtained and may help to discover not only the problem (a concept is difficult for students) but also its cause (maybe it is not explained in the materials of the course). The work presented in the paper proposes a novel approach to work in that direction: gathering information from different aspects within subjects. In particular, the paper presents an approach that uses natural language processing techniques to analyze the subject's materials in order to discover which concepts are taught and their importance in the subject. The contribution of the paper is a system that allows obtaining a better understanding of subjects. The results can be used for analyzing the suitability of materials to subjects and to enrich and contextualize other analytical processes.