Modelo e framework para o desenvolvimento de ferramentas analíticas de apoio ao ensino, aprendizagem e gestão educacional

• Ano de defesa: 2014
• Autor(a) principal: Rosales, Gislaine Cristina Micheloti
• Orientador(a): Araujo, Regina Borges de
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de São Carlos
• Programa de Pós-Graduação: Programa de Pós-Graduação em Ciência da Computação - PPGCC
• País: BR
• Palavras-chave em Português: Ciência da computação; Educação a distância; Análise da aprendizagem; Análise de dados; Educação - processo decisório; Ciência de contexto; Tomada de decisão educacional
• Palavras-chave em Inglês: Distance education; Learning analysis; Analysis of educational data; Educational decision making; Context awareness
• Área do conhecimento CNPq: CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: https://repositorio.ufscar.br/handle/20.500.14289/294

The use of new information and communication technologies in education that go beyond the traditional Learning Management Systems (LMS), has generated a growing volume of data, making challenging and complex the analysis of data generated to meet the decision-making levels of teaching, learning and management. Despite high expectations for data analysis in education, current research in the area is focused more specifically on student data, processes and learning behaviors, even when the focus of the research is to improve the teaching or actions at the institutional level. In order to facilitate and extend the process of data analysis in the areas of teaching, learning and management for different stakeholders, this thesis presents a conceptual model that will guide the construction of educational analytical context aware applications that support educational decision making in the micro, meso and macro levels. The conceptual model proposes the collection of educational data from multiple, heterogeneous sources in a decentralized manner using logical and physical sensors. The Model supports analysis of data collected at three levels: descriptive analysis, predictive analysis and prescriptive analysis. The conceptual model was established from an open architecture framework, extensible and reusable, which offers a simpler and unified path for both the acquisition of user behaviors in online learning, and the modeling and analysis of the collected contexts. To validate the proposed conceptual model, three applications were developed, namely: ViTrackeR, to support self-regulated learning by providing visualization of data tracking and personalized recommendations; ViMonitor to support real-time, teams of teaching and academic management providing important information on students and tutors; and ViAssess, which provides support for secure assessments online. The conceptual model was evaluated and validated in a real environment (students, tutors, teachers and administrators). The framework was rated by both developers of educational and analytical tools and by expert researchers in the field of this research, obtaining very positive results. Evaluation results indicate that the proposed conceptual model supports the development of educational applications in the three analytical levels of decision making, micro, meso and macro, and also supports the three levels of analysis provided: descriptive, predictive and prescriptive.