Classificação de Aprendizes através de Mapas Auto-Organizáveis
Resumo
Resumo: A habilidade de elaborar modelos de aprendiz é um componente crítico nos sistemas tutores inteligentes. Uma das grandes dificuldades encontradas na construção de sistemas que se adaptem às caractrísticas do aprendiz é determinar os conhecimentos e crenças do aluno antes, durante e ao término das sessões de treinamento. Neste experimento, um questionário de diagnóstico inicial foi aplicado a 88 estudantes que iniciaram o curso de tecnologia em informática. Os dados obtidos foram interpretados através de Redes Neurais Artificais empregando o algoritmo de Kohonen e confrontados com a opinião do professor da disciplina após um mês de aula.
Abstract: The ability to formulate apprentice's models is a critical component in the Intelligent Tutor System. One of the great difficulties found in the construction of systems that it adapt to the apprentice's characteristics is to determine the knowledge and the apprentice's believes before, during and at the end of the training sessions. Several technologies can be used for the making of student models, Artificial Neural Nets (ANN) it is the subject of this work. The algorithm of Kohonen has been obtaining success as classifier of objects with attributes and values in common. In this experiment, a questionnaire of initial diagnosis was applied 88 new computer science students. The obtained data were interpreted through the self-organizing maps (Kohonen) and confronted with the teacher's opinion after a month of class.
Abstract: The ability to formulate apprentice's models is a critical component in the Intelligent Tutor System. One of the great difficulties found in the construction of systems that it adapt to the apprentice's characteristics is to determine the knowledge and the apprentice's believes before, during and at the end of the training sessions. Several technologies can be used for the making of student models, Artificial Neural Nets (ANN) it is the subject of this work. The algorithm of Kohonen has been obtaining success as classifier of objects with attributes and values in common. In this experiment, a questionnaire of initial diagnosis was applied 88 new computer science students. The obtained data were interpreted through the self-organizing maps (Kohonen) and confronted with the teacher's opinion after a month of class.
Texto completo:
PDFDOI: https://doi.org/10.5753/cbie.wie.2003.22-29