A Linguistic Modeling Approach to Characterize Items in Computarized Adaptive Test for Intelligent Tutor Systems Based on Competency

摘要

An Intelligent Tutor System based on Competency education (ITS-C) aims to personalize teaching processes according to student's competency profile and learning activities by means of artificial intelligence (AI) techniques. One of the most challenging process in ITS-C is the diagnosis process, so far it has been carried out by computerized adaptive tests (CAT) based on item response theory (IRT), in spite of the good performance, its construction requires a hard statistical calibration of a huge bank of items. Such processes are usually intractable in small institutions. To overcome previous difficulties, enhance the accuracy of diagnosis, and the adaptation to student's competence level this contribution proposes the use of teachers' knowledge to replace statistical calibration by modeling such expert's knowledge linguistically using the fuzzy linguistic approach.