A carregar...
Publicação
Outlier detection and characterization for students
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 1.17 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
Nowadays, web-based educational systems are being used as an essential tool to the learning
process and with them the universities are capable to collect data from the students and build
studies to understand their behavior. In these studies, the outliers’ students are not the main topic,
or they are removed because of their extreme behavior, so this thesis focuses on detecting the
outliers and understanding their behavior. Focusing on these students a methodology is proposed to
find two clusters using outliers’ techniques on their grades and Moodle usage, the main goal was to
detect the subjects with a high volume of usage in Moodle and with a bad performance on the final
grade, this analysis was done for each subject that the student did. In the end we detect a total of
five students that require an intervention by the university to understand how their performance
could improve or if they are having some troubles in the use of the online tool. The methodology
proposed shows an efficient and easy way to find the students and could be replicated easily for
other universities and be implemented as an active tool to assist the university.
Descrição
Dissertation presented as the partial requirement for obtaining a Master's degree in Data Science and Advanced Analytics, specialization in Data Science
Palavras-chave
Outlier 1 Educational data mining 2 Moodle 3 SDG 4 - Quality education SDG 9 - Industry, innovation and infrastructure