In the fast-paced and competitive world of banking, customer retention has emerged as a crucial factor for sustainable growth and success. Beta Bank, a prominent financial institution, has been facing a gradual exodus of customers, as they chip away little by little each month. Recognizing the significance of preserving existing clients, Beta Bank has embarked on a mission to predict customer churn and implement strategic measures to retain valued customers. This project aims to develop a robust churn prediction model using historical customer behavior data and contract termination records to identify customers who are likely to leave the bank soon.
The project's primary goal is to develop a churn prediction model with the maximum possible F1 score to accurately predict whether a customer will leave Beta Bank soon. To pass the project, the model needs to achieve an F1 score of at least 0.59 on the test set, indicating a strong predictive performance in correctly identifying customers who are likely to churn. The secondary goal is to assess the model's performance using the AUC-ROC metric and compare it with the F1 score to gain comprehensive insights into the model's predictive capabilities.