Automarking: Automatic Assessment of Open Questions

A number of Learning Management Systems (LMSs) exist on the market today. A subset of a LMS is the component in which student assessment is managed. In some forms of assessment, such as open questions, the LMS is incapable of evaluating the students´ responses and therefore human intervention is necessary. In order to assess at higher levels of Bloom´s (1956) taxonomy, it is necessary to include open-style questions in which the student is given the task as well as the freedom to arrive at a response without the comfort of recall words and/or phrases. Automating the assessment process of open questions is an area of research that has been ongoing since the 1960s. Earlier work focused on statistical or probabilistic approaches based primarily on conceptual understanding. Recent gains in Natural Language Processing have resulted in a shift in the way in which free text can be evaluated. This has allowed for a more linguistic approach which focuses heavily on factual understanding. This study will leverage the research conducted in recent studies in the area of Natural Language Processing, Information Extraction and Information Retrieval in order to provide a fair, timely and accurate assessment of student responses to open questions based on the semantic meaning of those responses.