PRODUÇÃO ACADÊMICA Repositório Acadêmico da Graduação (RAG) TCC Engenharia de Computação
Use este identificador para citar ou linkar para este item: https://repositorio.pucgoias.edu.br/jspui/handle/123456789/10540
Registro completo de metadados
Campo DCValorIdioma
dc.creatorCosta, Arthurpt_BR
dc.date.accessioned2026-06-17T13:20:34Z-
dc.date.available2026-06-17T13:20:34Z-
dc.date.issued2026-06-12-
dc.identifier.urihttps://repositorio.pucgoias.edu.br/jspui/handle/123456789/10540-
dc.languageengpt_BR
dc.publisherPontifícia Universidade Católica de Goiáspt_BR
dc.rightsAcesso Abertopt_BR
dc.subjectMachine Learning, Churn prediction, Payment fintech.pt_BR
dc.titleAnálise preditiva de churn em uma fintech de pagamentos usando aprendizagem de máquina e redes neuraispt_BR
dc.typeTrabalho de Conclusão de Cursopt_BR
dc.contributor.advisor1Vinhal, Gustavo Siqueirapt_BR
dc.contributor.advisor1Latteshttp://lattes.cnpq.br/5227400971565575pt_BR
dc.contributor.referee1Abadia, Fernando Gonçalvespt_BR
dc.contributor.referee1Latteshttp://lattes.cnpq.br/3382052342707576pt_BR
dc.contributor.referee2Amaral, Nilson Cardosopt_BR
dc.contributor.referee2Latteshttp://lattes.cnpq.br/6824122529171550pt_BR
dc.description.resumoIn payment fintechs, merchant retention is closely linked to processed volume sustainability and reacquisition costs. Unlike subscription-based services, churn rarely involves explicit cancellation and must be inferred from transaction dynamics affected by seasonality and operational fluctuations. This study frames churn prediction as a time-oriented supervised binary classification problem using merchant-level aggregated transactional time series. A dual labeling scheme combines hard churn (30 consecutive days without transactions) with imminent churn (sustained volume decline relative to a historical baseline), serving as an early-warning proxy for relationship deterioration. Temporal features from rolling windows of 7, 14, and 28 days capture recency, intensity, trend, and variability. A time-based train–test split and walk-forward cross-validation reduce leakage under non-stationarity. On the holdout set, XGBoost achieved ROC-AUC of 0.830, PR-AUC of 0.655 (≈2.02× above random), and Recall@10% of 0.223 (≈2.2× random). Calibration and interpretability analyses, including permutation importance and Shapley values, support threshold definition and businessoriented decision-making.pt_BR
dc.publisher.countryBrasilpt_BR
dc.publisher.departmentEscola Politécnicapt_BR
dc.publisher.initialsPUC Goiáspt_BR
dc.subject.cnpqCNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO::SISTEMAS DE COMPUTACAOpt_BR
dc.degree.graduationEngenharia de Computaçãopt_BR
dc.degree.levelGraduaçãopt_BR
Aparece nas coleções:TCC Engenharia de Computação

Arquivos associados a este item:
Arquivo TamanhoFormato 
Mesclado.pdf
  Until 2200-01-01		1,52 MBAdobe PDFVisualizar/Abrir    Solictar uma cópia
Mostrar registro simples do item Visualizar estatísticas


Os itens no repositório estão protegidos por copyright, com todos os direitos reservados, salvo quando é indicado o contrário.

Ferramentas do administrador