Publicação
How machine learning can improve customer loss prevention: A case study in the banking sector
| datacite.subject.fos | Ciências Naturais::Ciências da Computação e da Informação | pt_PT |
| dc.contributor.advisor | Bação, Fernando José Ferreira Lucas | |
| dc.contributor.author | Macedo, Miguel José dos Santos Moita de | |
| dc.date.accessioned | 2025-11-12T09:29:38Z | |
| dc.date.available | 2025-11-12T09:29:38Z | |
| dc.date.issued | 2025-10-28 | |
| dc.description | Dissertation presented as the partial requirement for obtaining a Master's degree in Information Management, specialization in Business Intelligence | pt_PT |
| dc.description.abstract | The banking industry faces mounting pressure to retain profitable clients, as acquiring new ones is far costlier than preserving existing relationships. This study investigates how machinelearning–driven churn prediction can bolster customer-loss prevention for a Portuguese retail bank. A proprietary dataset containing 10 000 customers and 141 behavioural, transactional and demographic variables was explored under the CRISP-DM framework. After rigorous data cleaning, feature engineering, class-imbalance handling (SMOTE) and multi-stage feature selection, seven supervised models were benchmarked: Logistic Regression, Decision Tree, Random Forest, Gradient Boosting, Extreme Gradient Boosting (XGBoost), Support Vector Machine and Artificial Neural Network. Initial experiments showed ensemble techniques consistently outperforming single learners across Accuracy, Precision, Recall, ROC AUC and F1-score. Subsequent grid-search hyper-parameter tuning further improved results, with the tuned XGBoost model emerging as the top performer (F1 = 0.702; Accuracy = 0.800). Key predictive drivers included client tenure, recent debit-card activity, average deposit balances and digitalchannel engagement. These insights enable the bank to identify high-risk customers early and deploy tailored retention levers, such as repricing, loyalty incentives or personalised product bundles, before attrition occurs. Overall, the study demonstrates that a well-governed machine-learning pipeline can transform raw banking data into actionable intelligence, delivering a scalable decision-support tool that underpins proactive, data-driven churn-mitigation strategies. | pt_PT |
| dc.identifier.tid | 204071364 | |
| dc.identifier.uri | http://hdl.handle.net/10362/190559 | |
| dc.language.iso | eng | pt_PT |
| dc.rights.uri | http://creativecommons.org/licenses/by/4.0/ | pt_PT |
| dc.subject | Banking Sector | pt_PT |
| dc.subject | Customer Churn | pt_PT |
| dc.subject | Machine Learning | pt_PT |
| dc.subject | Predictive Models | pt_PT |
| dc.subject | SDG 8 - Decent work and economic growth | pt_PT |
| dc.subject | SDG 9 - Industry, innovation and infrastructure | pt_PT |
| dc.title | How machine learning can improve customer loss prevention: A case study in the banking sector | 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 Business Intelligence | pt_PT |