«Learning analytics, user modeling and adaptation in e-learning contexts»

Abstract

In the context of e-learning and e-training environments, Rosa M. Carro will address, firstly, the need and difficulty of building user models to store information about the different students so that it is feasible to adapt strategies, activities, and resources to each of them. Obtaining this information is not easy: sometimes it requires a great effort or it is obtained intrusively. In the first part of this talk she will present some works aimed at acquiring information as automatically and less intrusively as possible.

Once the students are being trained in the corresponding environments/applications, information about their interactions is also very valuable, so it must be stored in order to be analyzed, with a dual purpose: to re-feed user models, enriching them with data that reflect the student performance and progress; and to detect potential possibilities for improving the resources offered. Moreover, by using machine-learning algorithms it is possible to predict potential situations such as student failure or dropout, which allows carrying out in-time interventions to prevent them from happening. 

In the second part of this talk, Rosa M. Carro will describe the recent experience with SPA, an Abandonment Prediction System developed and put into production in a spanish online university. SPA has calculated hundreds of thousands of risk values, causing thousands of interventions aimed at retaining students. She will share with the audience the challenges and lessons learned during the transition from experimentation in the laboratory to production in a real university environment where teaching is conducted entirely online.

Bio

Rosa M. Carro got a PhD on Computer Science from Universidad Autónoma de Madrid (UAM) in 2001. She is professor of the Department of Computer Science at UAM since 2008. She visited the Universidade de Aveiro (2001-2002) and the Technische Universität München (2002-2003) to do research on adaptive educational games and collaborative adaptive e-learning systems. Currently she is interested in automatic user model acquisition, sentiment and social network analysis, adaptive e-learning systems, attention to diversity, educational data mining and learning analytics. She has co-organised several international workshops, edited special issues in international journals and coauthored around 100 papers in journals and conferences dealing with her main research areas. Mrs. Carro is member of the director board of the Association for the Development of Educational Informatics (ADIE).