Client Behavior Prediction Using AI in Banking

This project aims to predict how banking clients are likely to respond to direct marketing efforts. Using a combination of traditional machine learning algorithms, ensemble models, and neural networks, it analyzes client data to forecast whether a client will subscribe to a term deposit.

---

Project Goals

• Evaluate and compare various classification algorithms
• Build ensemble models to enhance prediction accuracy
• Integrate deep learning models using Keras
• Predict client actions based on past data
• Support banking strategies with data-driven insights

---

Data Source

The dataset is sourced from the UCI Machine Learning Repository. It includes client information gathered during a direct marketing campaign by a Portuguese bank.

Please cite if used:

Moro, S., Cortez, P., & Rita, P. (2014). A data-driven approach to predict the success of bank telemarketing. Decision Support Systems, 62, 22–31.

---

Techniques Used

Machine Learning Models:

• Logistic Regression
• Quadratic Discriminant Analysis
• Naive Bayes
• Random Forest
• AdaBoost

Ensemble Approaches:

• Soft Voting Classifier
• Bagging with Neural Networks
• Stacked Deep Learning Models

Deep Learning:

• Custom Neural Network using Keras
• Integration of Keras with scikit-learn
• Ensemble learning with multiple neural nets

---