Computer adaptive testing for an intelligent E-learning platform

Abstract. This Thesis proposes a Computer Adaptive Test model based on Item Response Theory for an E-Learning platform. The Computer Programming course of Universidad Nacional de Colombia is used to proof the concept of this platform. In order to integrate the model, two modules that allow the creation and administration of tests are developed: test management module and test taking module. Every Computer Adaptive Test, uses an item bank to select items according to one of the Item Response Theory models available in the platform (One, Two and Three Parametric Logistic models). Each item bank is created using the experts approach, with the collaboration of professors of the course. The development of the Computer Adaptive Test model is described, including, its versions, trials, and how it is improved using feedback from students and professors. In addition to this, two data mining techniques are used in order to establish users prfoiles and question associations. These processes uses data that is generated after each trials and uses the behavior of students (their answers) in order to generate useful information for the course. The first data mining process, generates three cluster of students per test using the k-means algorithm, these clusters are analyzed in order to generate a proper description. The resulting descriptions shows that students are group according to its ability level. In the second data mining process the associations rules process generates relation between questions using the Apriori algorithm. There relations are analyzed and reveals relations between topics of the course. This information can be used to review the structure of the course content based on students performance.

• Autores: Mahecha D´Maria, Julián Ricardo
• Fecha de publicación: 2014
• Institución: Universidad Nacional de Colombia
• Repositorio: Universidad Nacional de Colombia
• Idioma: spa
• OAI Identifier: oai:repositorio.unal.edu.co:unal/51996
• Acceso en línea: https://repositorio.unal.edu.co/handle/unal/51996; http://bdigital.unal.edu.co/46244/
• Palabra clave: 0 Generalidades / Computer science, information and general works; 37 Educación / Education; 62 Ingeniería y operaciones afines / Engineering; E-Learning; Virtual Platform; Computer Adaptive Tests; Item Response Theory; Data mining; Aprendizaje en Línea; Plataforma Virtual; Pruebas Adaptativas por Computador; Teoría de Respuesta al Item; Minería de Datos
• Rights: openAccess
• License: Atribución-NoComercial 4.0 Internacional