Predicting student performance in foreign languages with a serious game

Abstract

In this digital age, many statements have been made regarding the use of technology for teaching purposes. In this sense Serious Games are gaining ground considering that, besides their technological advantages, they provide fun, which allegedly engages students in their training. Much research has been carried out to show how Serious Games improve teaching methodologies and student learning outcomes in various subjects. This research focuses on the field of digital game-based learning from a different perspective: Namely, the work carried out does not focus on the use of Serious Games for teaching and learning, but on the use of such tools for the prediction of learning outcomes. Accurately predicting future student performance lets teachers give customized advice to them. The approach is undertaken by means of machine learning and data mining techniques, and educational data mining techniques in particular. These techniques are applied to data collected from games played by students. For such purposes, The Conference Interpreter (CoIn), a Serious Game which simulates a context of simultaneous interpreting has been developed and used as a data mining tool. Following this, the experiment carried out is described and machine learning/data mining results are presented and discussed.