«Improving the evaluation of serious games by applying learning analytics and data mining techniques»

Abstract

Serious games are motivating and effective resources to teach, raise awareness, or change the perceptions of your players. The multiple interactions that players make can provide accurate information to assess their learning. In this sense, learning analytics propose techniques for collecting and analyzing these interactions for multiple purposes, including the evaluation of the players. Data from learning analytics can be analyzed with data mining techniques that provide better measures for studying the effect of games on students and assessing their learning.

In this talk, we present a method for evaluating serious game players based on evidence collected while playing. The collected interaction data is analyzed to extract learning analytics variables used by predictive models to quantify player learning. The proposal uses a standard data collection format for interactions with serious games, xAPI-SG, which allows systematizing both the interaction data collected and the creation of learning analytics variables. The proposal has been based on two case studies with different serious games, which have made it possible to improve and generalize the process.

Bio

Cristina Alonso Fernández has a PhD in Computer Engineering from Universidad Complutense de Madrid. She has been part of the eUCM research group since September 2016, in which she has worked as a Contracted Researcher as part of the H2020 Beaconing project and collaborating in various national and European projects. Among her research interests are the study of educational video games and the application of analysis techniques and data mining for their improvement.