Abstract: Techniques are described by which learner skill can be estimated over time, even in the presence of large data sets. A rating system can be used in which a learner's submission of a course assessment can be interpreted as a single match, where the learner and the course assessment are competing against each other. This approach can be applied to online learning platforms that include very many learners who have submitted assignments very many times in many courses. The described techniques provide reliable estimates of assessment difficulty and learner ability. The described scoring framework may be used as a basis for various applications that account for a learner's ability, such as adaptive diagnostic tests and personalized recommendations.

Abstract: Techniques are described by which learner skill can be estimated over time, even in the presence of large data sets. A rating system can be used in which a learner's submission of a course assessment can be interpreted as a single match, where the learner and the course assessment are competing against each other. This approach can be applied to online learning platforms that include very many learners who have submitted assignments very many times in many courses. The described techniques provide reliable estimates of assessment difficulty and learner ability. The described scoring framework may be used as a basis for various applications that account for a learner's ability, such as adaptive diagnostic tests and personalized recommendations.