A personalized e-learning model for restricted social networks

Resumen

Social networks focused on a specific topic or community are a powerful and precise means for user communication and interconnectivity, no matter the role they stand for. These can be learners, teachers, employees, staff, academic managers, or financial directors, who show a very determined attitude, depending on their context and their objectives. Every user can question, answer, start an activity, follow another, comment on someone else¿s job, score a job made by others, search onto Internet, follow a scheduled test, participate in a video-conference with a teacher, and so on. And, in all these activities, any user can be pro-active, reactive, passive, consumer, producer, dealer, and yet to show some additional facets. This research analyses a learning ecosystem, restricted by user access and topic, and by the definition and implementation of a personalised learning model which deals with every single input and feature aforementioned, in order to provide the user with adaptive tutoring, thanks to a rule system. In this ecosystem, the users interact one with each other, and with the system, and they get personalised counselling. We develop a conceptual model, L.I.M.E. as for Learning, Interaction, Mentoring, Evaluation. These four vectors are measured and analysed as the pillars for the learning scenario, and they are depicted in various inputs which feed the model. The learning itinerary provided by the L.I.M.E. model increases the user performance. To prove it, we have design and implemented a learning scenario in a real class, which we have split in two groups (experimental and control) of 24 students, each. We have selected and analysed a subject of an official university online programme, during 4 weeks. This scenario engaged formal and informal activities with a comprehensive approach. The implementation shows successful results which prove the validity of the model. In addition, we have got useful recommendations and promising conclusions for further versions of the model. The combination of 48 learners, along 4 weeks and related milestones, the measurement of 30 inputs focused on informal and formal settings and distributed along the four main vectors, has resulted in a large dataset with sufficient information to retrieve meaningful and significant interpretation. The main outcome highlights that there is a clear and positive influence in the user performance, when the L.I.M.E. model is implemented. This conclusion is supported by a 10,53% overall average difference between the experimental group and the control group (66,72% - 56,19%), with a peak difference between corners of 37,37% (81,41% - 44,04%). These overall results, along with the partial ones which are presented along the dissertation, support seamlessly the online personalised learning model for thematic, restricted social networks, L.I.M.E.