Full Program »

Avaliação Automática em questões com diferentes graus de dificuldade: Uma abordagem utilizando K- Nearest Neighbor

This article presents a proposal for a virtual environment for Concept Maps in which there is a system of evaluation that focuses on two approaches: i) automatic assessment and ii) questions with variable difficulty degrees. For automatic assessment, it was programmed an algorithm that tests the isomorphism which returns various attributes which are input to a process of KNN that seeks a final score from a list of Concept Maps that was previously evaluated.

Author(s):

Márcio Pinheiro de Aragão    
Universidade Federal do Pará
Brazil

Eloi Luiz Favero    
Universidade Federal do Pará
Brazil