Loading...
Publication
MILAGE Aprender +: aprendizagem personalizada suportada por aprendizagem máquina
| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 3.21 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
O MILAGE Aprender+ é uma aplicação para dispositivos móveis com propósito educacional dentro e/ou fora do ambiente escolar. Os seus utilizadores têm acesso a uma ampla biblioteca de conteúdo para a realização de atividades diversas disciplinas escolares. Destaca-se por ser uma aplicação que prioriza a autonomia do aluno. Apresenta-se como uma ferramenta essencial para responder à atual procura relacionada com o ensino remoto e também presencial A aplicação MILAGE Aprender+ adota também uma característica jogo, contabilizando pontuações ao longo das realizações dos alunos. Desde o seu lançamento em 2018, possui um amplo conjunto de dados, um amplo histórico de atividades realizadas pelos seus utilizadores. No modelo atual de funcionamento o MILAGE Aprender+ é percebido um problema relacionado a forma que apresenta o conteúdo aos utilizadores. O problema é percebido pela forma que o conteúdo da aplicação é apresentado ao aluno, de forma totalmente uniforme e padronizada a todos os seus utilizadores, não tendo em consideração o contexto histórico do desempenho dos seus utilizadores ao longo do uso da aplicação. Não é aplicado nenhum tipo de recomendação de conteúdo ou apresentação de forma personalizada. Pelo que, o presente estudo, tem como objetivo avaliar a possibilidade de evolução da aplicação MILAGE Aprender+ com a adoção da tecnologia de Aprendizagem Máquina (AM) a fim de proporcionar uma melhor experiência ao seu utilizador, o aluno. Pretende-se recomendar conteúdos a partir de modelos que promovam melhorar o aproveitamento de pontos do aluno. A partir deste objetivo, o presente estudo aprofundou-se na AM, que é um método de análise e ciência de dados que automatiza a construção de modelos analíticos. Esta tese adotou a metodologia de desenvolvimento de projetos de prospeção e análise em ciência de dados CRISP-DM. A investigação desenvolveu uma ampla análise sobre o problema do MILAGE Aprender+, e a possibilidade da adoção da tecnologia de AM. O principal objetivo de permitir que os alunos, utilizadores da ferramenta, recebam a recomendação de conteúdos das disciplinas de forma adaptável, aproximada às suas necessidades e baseada no histórico de desempenho, para que assim possam assim ter um percurso de aprendizagem mais objetivo e também agradável. Como conclusão deste estudo evidenciou-se a possibilidade de adoção de modelos de AM no MILAGE Aprender+ para a identificação e recomendação de conteúdos aos alunos
MILAGE Learn+ is an application for mobile devices with educational purposes inside and/or outside the school environment. Its users have access to a wide library of content for carrying out activities in different school subjects. It stands out for being an application that prioritizes student autonomy. It is presented as an essential tool to respond to the current demand related to remote and face-to-face teaching. The MILAGE Learning+ application also adopts a game feature, counting scores along with the students' achievements. Since its launch in 2018, it has had a wide set of data, a wide history of activities carried out by its users. In the current operating model of MILAGE Learning+, a problem is perceived related to the way it presents the content to users. The problem is perceived by the way that the content of the application is presented to the student, in a totally uniform and standardized way to all its users, not considering the historical context of the performance of its users throughout the use of the application. No content recommendation or personalized presentation is applied. Therefore, the present study aims to evaluate the possibility of evolving the MILAGE Learning+ application with the adoption of Machine Learning (ML) technology to provide a better experience to its user, the student. It is intended to recommend content based on models that promote the improvement of the student's use of points. From this objective, the present study delved into AM, which is a data science and analysis method that automates the construction of analytical models. This thesis adopted the CRISP-DM data science prospecting and analysis project development methodology. The investigation developed a broad analysis of the problem of MILAGE Learn+, and the possibility of adopting ML technology. The main objective is to allow students, users of the tool, to receive the recommendation of contents of the disciplines in an adaptable way, close to their needs and based on the performance history so that they can thus have a more objective and pleasant learning path. As a conclusion of this study, the possibility of adopting ML models in MILAGE Learn+ was evidenced to identify and recommend content to students.
MILAGE Learn+ is an application for mobile devices with educational purposes inside and/or outside the school environment. Its users have access to a wide library of content for carrying out activities in different school subjects. It stands out for being an application that prioritizes student autonomy. It is presented as an essential tool to respond to the current demand related to remote and face-to-face teaching. The MILAGE Learning+ application also adopts a game feature, counting scores along with the students' achievements. Since its launch in 2018, it has had a wide set of data, a wide history of activities carried out by its users. In the current operating model of MILAGE Learning+, a problem is perceived related to the way it presents the content to users. The problem is perceived by the way that the content of the application is presented to the student, in a totally uniform and standardized way to all its users, not considering the historical context of the performance of its users throughout the use of the application. No content recommendation or personalized presentation is applied. Therefore, the present study aims to evaluate the possibility of evolving the MILAGE Learning+ application with the adoption of Machine Learning (ML) technology to provide a better experience to its user, the student. It is intended to recommend content based on models that promote the improvement of the student's use of points. From this objective, the present study delved into AM, which is a data science and analysis method that automates the construction of analytical models. This thesis adopted the CRISP-DM data science prospecting and analysis project development methodology. The investigation developed a broad analysis of the problem of MILAGE Learn+, and the possibility of adopting ML technology. The main objective is to allow students, users of the tool, to receive the recommendation of contents of the disciplines in an adaptable way, close to their needs and based on the performance history so that they can thus have a more objective and pleasant learning path. As a conclusion of this study, the possibility of adopting ML models in MILAGE Learn+ was evidenced to identify and recommend content to students.
Description
Keywords
Aprendizagem máquina Milage aprender+ Am automatizada Ensino à distância