O uso de mapas auto-organizáveis como ferramenta de análise exploratória em testes cognitivos destinados a medir o desempenho escolar

• Ano de defesa: 2017
• Autor(a) principal: Jarske, Johne Marcus
• Orientador(a): Silva, Leandro Augusto da
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Presbiteriana Mackenzie
• Palavras-chave em Português: análise exploratória; visualização de dados; mapas auto-organizáveis; estatística inferencial; testes cognitivos; desempenho acadêmico; desenvolvimento cognitivo; aprendizado da leitura
• Área do conhecimento CNPq: CNPQ::ENGENHARIAS
• Link de acesso: http://dspace.mackenzie.br/handle/10899/24457

Despite the improvement in schooling rates pointed by international indicators such as PISA 2012/2015 - Program for International Student Assessment, the education quality in Brazil continues to be very poor, necessitating improvements in practically all aspects assessed by the OECD, which organizes PISA, such as performance in reading and mathematics. This study aims to contribute with techniques that can improve the analysis of educational data and proposes the use of self-organizing maps (SOM) as a tool for exploratory data analysis to support discoveries and diagnoses related to school performance, focusing on the analysis of tests developed to measure the development of cognitive abilities in students. It is hoped that these techniques of analysis can help educational researchers in the preparation of better diagnoses related to student’s cognitive development and to assist in the validation process and normalization of the cognitive tests, thus providing techniques for data visualization, patterns and outliers detection, cluster analysis and searching at hidden information in the data.