Núcleos de avaliações diagnóstica e formativa para regulação da aprendizagem de programação

• Autor(a) principal: Oliveira, Márcia Gonçalves de
• Data de Publicação: 2013
• Tipo de documento: Tese
• Idioma: por
• Título da fonte: Repositório Institucional da Universidade Federal do Espírito Santo (riUfes)
• Texto Completo: http://repositorio.ufes.br/handle/10/9711
• topic: Avaliação Diagnóstica; Avaliação Formativa; Avaliação Semi-automática de Exercícios; Recomendação de Atividades, Aprendizagem de Programação; Engenharia Elétrica; Avaliação; Aprendizagem; Microcomputadores - Programação; Reconhecimento de padrões; 621.3

The computers programming knowledge is considered complex because it requires a combination of several cognitive skills and extensive practice to be learned. Once many students come to universities malformed in essential skills to solve problems, it is expected that they face difficulties to develop computer programs. Thus, programming disciplines are those which respond more for the high failure rates and even evasion of computers science courses. As the programming classrooms are more numerous every year by increasing the number of vacancies and the additional students who failed, manual and cognitive effort is demanded from the teacher to correct a lot of exercises. Furthermore, in classrooms with a lot of students, it is almost impossible for a teacher to make an individual learning monitoring of their students. In order to reduce teacher’s efforts to correct exercises and to provide better learning conditions for students of programming courses, we propose in this thesis a system for monitoring and regulation of learning programming. This system consists of two cores: Diagnostic Assessment Core (DAC) and Formative Assessment Core (FAC), both implemented by pattern recognition technologies such as Clustering, Linear Regression and ML-kNN Multilabel Classification Algorithm. The DAC performs the functions of semi-automatic correction of exercises and mapping of students’profile. The FAC, in turn, has the functions of stability control of student’s performances and of activities recommendation for students whose profiles indicate learning difficulties. In this thesis the assessment cores were applied in the learning context of C Language. The results of applying assesment cores in actual programming classrooms demonstrate that it is possible automatically to reduce the correction effort of exercises up to 70% and to mimic the recommendations of teachers around 90% of the times. In summary, the contribution of this work to the computer programming field is to provide a mechanism for diagnosis and adjustment of variables that characterize the learning programming allowing the teacher to make better learning management of their students.