Dynamic student assessment to advocate personalized learning plan

A central challenge in education is to match instruction to the characteristics and learning styles of students in order to optimize learning. In this article, we intend to outline our approach to supporting personalized learning strategies by constructing dynamical student profiling using ubiquitous computing capability. This profiling includes recorded data on students´ affective responses to learning to discern students´ level of motivation and details from generic student profiles to describe and predict student learning patterns. Learning pattern data analysis derives conclusions using decision trees. Through this process, information can be extracted from students´ affective responses and students´ profile data and relevant correlations between the two data sets can be recognized automatically. A personalized learning component uses this information to offer proactive support to students. This is achieved by recommending personalized courses of action which are beneficial to students. Our proposed model has been tested in a classroom simulation. Issues of sample limitations and promising directions for future research are elaborated towards the end of this paper.