Publicação
Predicting the acceptance of students in NOVA IMS Post-graduate´s courses to support decision-making in the university admission process
| datacite.subject.fos | Ciências Naturais::Ciências da Computação e da Informação | pt_PT |
| dc.contributor.advisor | Henriques, Roberto André Pereira | |
| dc.contributor.advisor | Oliveira, Lara Barradas Teixeira Garrucho de | |
| dc.contributor.author | Neta, Maria Zeneide Mota Veras | |
| dc.date.accessioned | 2024-03-01T17:50:18Z | |
| dc.date.available | 2024-03-01T17:50:18Z | |
| dc.date.issued | 2024-01-31 | |
| dc.description | Dissertation presented as the partial requirement for obtaining a Master's degree in Information Management, specialization in Knowledge Management and Business Intelligence | pt_PT |
| dc.description.abstract | In the data-driven educational context, this study focuses on supporting the admissions decision-making for all post-graduate programs offered by NOVA IMS. Employing the CRISP-DM methodology and a diverse dataset, including structured and unstructured candidate data, student performance records, and external university rankings, the research creates two distinct models. The first model, based on classifier techniques, is designed to predict the candidate acceptance. The second model, as a complementary model, applies regression techniques to predict the candidate’s performance. The regression model aims to validate the selection of top candidates at the conclusion of the admission process by providing a list of candidates predicted to perform well. Both models are designed to take advantage of the available historical university data and enhance the manual and time-consuming admission process performed every year by the university's selection committee. Three text mining techniques (TF-IDF, Word2Vec and Doc2Vec) were applied to the unstructured data, which included the candidate’s motivation statement, as well as textual information about their past courses and work experiences. The modelling algorithms evaluated in this study for both models were Artificial Neural Network, Decision Tree, Random Forest, and Support Vector Machine. The Random Forest Classifier delivered superior predictions for the first model, whereas the Support Vector Regression performed better for the second model. Finally, both models produced satisfactory results, successfully achieving the study’s objective. | pt_PT |
| dc.identifier.tid | 203531361 | pt_PT |
| dc.identifier.uri | http://hdl.handle.net/10362/164329 | |
| dc.language.iso | eng | pt_PT |
| dc.rights.uri | http://creativecommons.org/licenses/by/4.0/ | pt_PT |
| dc.subject | Educational Data Mining | pt_PT |
| dc.subject | Machine Learning | pt_PT |
| dc.subject | Admission prediction | pt_PT |
| dc.subject | Performance prediction | pt_PT |
| dc.subject | Decision support | pt_PT |
| dc.subject | SDG 4 - Quality education | pt_PT |
| dc.title | Predicting the acceptance of students in NOVA IMS Post-graduate´s courses to support decision-making in the university admission process | pt_PT |
| dc.type | master thesis | |
| dspace.entity.type | Publication | |
| rcaap.rights | openAccess | pt_PT |
| rcaap.type | masterThesis | pt_PT |
| thesis.degree.name | Mestrado em Gestão de Informação, especialização em Gestão do Conhecimento e Inteligência de Negócio (Business Intelligence) | pt_PT |