Recomendação de Estudos no Ambiente Virtual de Aprendizado do AlfaCon Concursos Públicos: estudos com uma API RESTful.

• Autor(a) principal: Malacarne, Gustavo Raí
• Data de Publicação: 2024
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Biblioteca Digital de Teses e Dissertações do UNIOESTE
• Texto Completo: https://tede.unioeste.br/handle/tede/7232
• topic: Sistema de Recomendação; Recomendação de Estudos; AfaCon; API; TCT; TRI; Recommendation System; Study Recommendation; AlfaCon; API; CTT; IRT; CIÊNCIA DA COMPUTAÇÃO

This work presents the development and initial studies of an Application Programming Interface (API) for the Virtual Learning Environment (VLE) of AlfaCon, focused on preparing students for public service exams. The API, based on the Theory of Meaningful Learning (TML), Classical Test Theory (CTT), and Item Response Theory (IRT), makes study recommendations to students who opt for its use and commit to providing additional personal data and answers from simulated exams. This is part of a broader effort by the company to transition from the traditional VLE currently in use to an adaptive VLE, which makes content recommendations based on its duly cataloged Educational Objects and their didactic and operational characteristics. The indication pointed out by this research for AlfaCon to conduct such a transition, especially due to the scarcity of complete and adequate data related to students and the actions they and the Pedagogical Team (PT) perform in the VLE, is the use of the API, for a minimum period of time, so that stakeholders have more and better convictions about the specifics and needs regarding the requirements of the intended adaptive VLE. The API operates independently of AlfaCon’s current VLE, not interfering with the dynamics of ongoing courses but collecting necessary data. During a simulated exam for the Federal Highway Police (PRF) preparatory course, made available by AlfaCon at the end of 2023, a real-time test with the API was conducted with 89 volunteer students, enabling an evaluation involving CTT metrics and respondents’ prior knowledge, as predicted by TML, to identify areas of knowledge where they showed greater difficulties and other aspects. IRT contributed to identifying discrimination parameters, difficulty, and the chance of guessing correctly on the items of the simulated exam, enabling the creation of charts to enhance analyses made by the PT. The evaluations conducted with the tests related to these 89 students pointed out that the API effectively makes study recommendations and also provides a customized individual report, with intuitive infographics and analytical metrics, facilitating a better understanding of the individual evolutionary trajectory throughout the course. The company’s PT also receives feedback that identifies students’ progress during the course, assesses the effectiveness of the items included in the simulations, and quantifies a test’s aptitude in measuring respondents’ skills. Of the students who evaluated the report, 68.2% gave the highest score (5), while 22.7% and 9.1% assigned scores 4 and 3, respectively. Moreover, 95.2% perceived the recommendations as beneficial for understanding and improving their skills in the indicated areas, evidencing the positive impact of the report on their self-assessment and study planning.