This work develops a method for incorporation into an onlinesystem to provide carefullytargeted guidance and feedback to students. The student answers onlinemultiple choice questions ona selected topic, and their responses are sent to a SnapDriftneural network trained with responsesfrom past students. Snapdriftis able to categorise the learner's responses as having a significant levelof similarity with a subset of the students it has previously categorised. Each category is associatedwith feedback composed by the lecturer on the basis of the level of understanding and prevalentmisconceptions of that categorygroupof students. In this way the feedback addresses the level ofknowledge of the individual and guides them towards a greater understanding of particular concepts.The feedback is conceptbasedrather than tied to any particular question, and so the learner isencouraged to retake the same test and receives different feedback depending on their evolving state ofknowledge. This approach has been applied to two data sets related to topics from an Introduction toComputer System module and a Research Skills module.
展开▼
