Belo, RodrigoGrade, MiguelCunha, António Templer da2025-02-242024-01-292024-01-29http://hdl.handle.net/10362/179656Plano +CUF is a subscription-based healthcare plan created by CUF Hospitais e Clínicas. The goal of this project is to create a model to predict client churn from this healthcare plan. A reliable predictive model can help CUF create actions to prevent churn and avoid clients from abandoning the service. If these actions are effective, then CUF can retain clients and prevent a potential loss. From model results and feature importance analysis, many “churning” clients could be detected and the main indicator that a client is going to churn is when it stops using the plan and its benefits.engChurnMachine learningPredictive modelHealthcareSubscription-basedmaster thesis203867491