Sistema Adaptativo de Questões do Enem Baseado em Teoria de Resposta ao Item

Abstract

This research addresses the challenge of personalized study for the ENEM (Brazilian National High School Exam) by developing an adaptive question recommendation system based on Item Response Theory (IRT). Students may struggle to select appropriate study questions, leading to frustration. This work involved the development of a prototype web application that recommends progressive ENEM questions, leveraging IRT to personalize the learning path. The prototype includes essential features such as authentication, initial proficiency assessment, adaptive recommendation, performance dashboard, and question review. The study included the building of a question database, implementing an IRT- based recommendation algorithm, and conducting a pilot study with 27 first-year high school students to assess user perception of utility and progress. The results validated the central hypothesis, revealing positive perceptions across all evaluated dimensions: high usability (mean 4.51), strong engagement (mean 4.35), and a recommendation rate of 92.6% (25 out of 27 students), indicating strong acceptance of the IRT-based approach.