Personalised collaborative skills for student models

Student models are crucial components in personalised distance learning environments. These models usually include individual characteristics such as the level of knowledge of a given topic, the learning style or the type of personality, the level of participation and so on. However, when the focus is on group activities, these learning environments often use group characteristics, losing sight of the individuals in the groups. Therefore, a student model considering individual characteristics in groups is introduced in this article. This model is built by automatically analysing collaborative behaviour and capturing the context in which it appears. A particular contribution of this approach is focused on a flexible structure of collaboration contexts which are defined at runtime. This approach has been proven in simulated and real environments showing promising results, provided that a given collaborative behaviour is found in at least 20% of the captured actions.