Semi-auto assessing on-line submitted texts with ontology-based structure map (ATOM)

This study proposes a semi-automatic procedure to assess text answers submitted by online learners. It tries to figure out the ontology description of the teacher toward the best answer for the question given. With the interactive procedure, the algorithm presents the structure of each text answer to the teacher, who may confirm or remedy it. Besides, the algorithm will give a pair of deliberately picked texts to the teacher, who then compares them and judges their goodness in rankings, which again are transformed to scores to learn the teacher´s ontology-based structure map. When the algorithm finishes capturing what the teacher is thinking about, it takes over the remaining assessments; hence alleviates the load of the teacher and raises the reliability, since all scores are given now with the same standard just learned from the teacher.