This academic project focuses on modeling churn (customer attrition rate) in a banking context. Utilizing Python and Machine Learning algorithms, the model analyzes client data to predict which ones are more likely to leave the bank’s services. The aim is to provide valuable insights that can be utilized for customer retention strategies.
The project was developed in the Python Colab environment, using libraries such as NumPy, Pandas, Scikit-learn, and Keras. The analysis begins with data pre-processing, where categorical variables like ‘Geography’ and ‘Gender’ are transformed using Label Encoding. The Random Forest Classifier model is trained with the data, achieving an accuracy of 86.83%, highlighting its effectiveness in predicting customer churn.
Users can input information such as credit score, geographic location, gender, age, account balance, among others, and the model uses this information to predict whether the customer is at risk of churn. In addition to predicting churn, the model also provides the associated probability of this prediction, allowing for a deeper understanding of customer behavior.