Knowledge discovery potential in automated test assembly

Creating knowledge tests in academic institutions is a procedure that is frequently repeated, and which requires examiner´s effort and significant time invested. The procedure is complex and demands objectivity of the examiner, test consistency and reliability. Automating the procedure can improve objectivity and facilitate the process to the examiner. This paper analyses potential of the proposed method for automated test assembly based on existing test bank, using method knowledge discovery and the nearest neighbor algorithm.