Unfolding the drivers for academic success: The case of ISCTE-IUL

• Autor(a) principal: Gil, Paulo Alexandre Vieira Diniz Ferreira
• Data de Publicação: 2019
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10071/20069
• topic: Academic success; Data mining; Modelling; SVM; Features; Sensitivity analysis; Sucesso escolar; Modelação; Características; Análise de sensibilidade; ISCTE Instituto Universitário de Lisboa; Ensino superior; Modelos de previsão

Predicting the success of academic students is a major topic in the higher education research community. This study presents a data mining approach to predict academic success in a Portuguese University called ISCTE-IUL, unveiling the features that better explain failures. A dataset of 10 curricular years for bachelor’s degrees has been analysed. Features’ selection resulted in a characterising set of 68 features, encompassing socio-demographic, social origin, previous education, special statutes and educational path information. Understanding features’ collection timings, distinct predicting was conducted. Based on entrance date, end of the first and the second curricular semesters, three distinct data models were proposed and tested. An additional model was designed for outlier degrees (i.e., a 4-year Bachelor). Six algorithms were tested for modelling. A support vector machines (SVM) model achieved the best overall performance and was selected to conduct a data-based sensitivity analysis. Relevance and impact review allowed extracting meaningful knowledge. This approach unfolded that previous evaluation performance, study gaps and age-related features play a major role in explaining failures at entrance stage. For subsequent stages, current evaluation performance features unveil their predicting power. Also, it should be noted that most of the features’ groups are represented on each model’s most relevant features, revealing that academic success is a combination of a wide range of distinct factors. These and many other findings, such as, age-related features increasing impact at the end first curricular semester, set a baseline for success improvement recommendations, and for easier data mining adoption by Higher Education institutions. Suggested guidelines include to provide study support groups to risk profiles and to create monitoring frameworks. From a practical standpoint, a data-driven decision-making framework based on these models can be used to promote academic success.