ChurnInsight AI: Predictive Analytics for Customer Retention in Banking 🚀

Overview 🧠

ChurnInsight AI is a machine learning project designed to predict customer churn in the banking sector. The project aims to identify at-risk customers early, enabling banks to implement proactive retention strategies. By leveraging behavioral and demographic data, ChurnInsight achieves accurate predictions and provides actionable insights to reduce churn rates.

Problem Statement 💡

Customer churn is a critical issue in the banking industry, directly impacting profitability. Retaining customers is more cost-effective than acquiring new ones, but traditional methods often fail to detect churn in time. This project addresses this gap using advanced machine learning techniques.

Features ✨

• Predict customer churn using machine learning algorithms.
• Provide insights into key factors influencing churn. 🔍
• Handle class imbalance using synthetic oversampling (SMOTE).
• Compare model performance using metrics like F1 Score and ROC-AUC.

Dataset 📊

The dataset, sourced from Kaggle, contains information on 10,000 customers from ABC Multinational Bank. It includes 14 attributes, such as:

• Demographic Data: Age, gender, etc.
• Behavioral Data: Balance, number of products, tenure, etc.
• Outcome: Whether the customer churned (binary classification).

Data Preprocessing 🛠️

• Data Cleaning: Checked for null values and duplicates.
• Outlier Handling: Managed extreme values to improve model performance.
• Class Balancing: Applied SMOTE to address class imbalance.
• Feature Engineering: One-hot encoding for categorical features and normalization for numerical features.

Machine Learning Pipeline 🏗️

Models Implemented

1. Logistic Regression: Simple baseline for binary classification.
2. k-Nearest Neighbors (k-NN): Uses customer similarity for predictions.
3. Decision Trees: Captures non-linear patterns and provides interpretability.
4. Random Forest: Reduces overfitting using ensemble learning. 🌳
5. XGBoost: High-accuracy model with effective handling of imbalanced data. ⚡

Evaluation Metrics 📈

• F1 Score: Measures the balance between precision and recall.
• ROC-AUC: Evaluates the model’s ability to distinguish between churned and non-churned customers.

Results 🏆

• Best Model: Random Forest
  • Validation F1 Score: 0.72
  • Validation ROC-AUC: 0.80
• Other models like Logistic Regression and Decision Trees showed competitive performance but were outperformed by Random Forest.

Challenges ⚠️

• Class Imbalance: Resolved using SMOTE.
• Feature Selection: Addressed using domain knowledge and exploratory data analysis.
• Model Interpretability: Enhanced through hyperparameter tuning and visualization.

Future Work 🔮

• Enhance predictive accuracy through advanced feature engineering and optimization.
• Integrate real-time data streams for dynamic churn predictions. ⏱️
• Explore neural networks and sentiment analysis to capture complex patterns and customer behavior.

How to Use 🛠️

1. Clone the Repository:

   git clone <repository-url>

2. Install Dependencies:

   pip install -r requirements.txt

3. Run the Notebook: Open the Main.ipynb file in Jupyter Notebook or any compatible IDE.

4. Train the Models: Follow the steps in the notebook to train and evaluate the models.

5. Analyze Results: Review the visualizations and metrics to understand model performance. 📊

Directory Structure 📁

ChurnInsight-AI/
├── data/                # Dataset and processed data files
├── notebooks/           # Jupyter notebooks for EDA and modeling
├── results/             # Visualizations and performance metrics
├── README.md            # Project overview and instructions
├── requirements.txt     # Python dependencies

Dependencies 🧩

• Python 3.x
• Libraries: pandas, numpy, matplotlib, seaborn, scikit-learn, xgboost

Authors ✍

• Aiswaryaa Velumani
• Sai Teja Aggunna
• Shrinisha Nirmalkumar
• Vaishnavi Pathipati

Acknowledgments

• Dataset Source: Kaggle

For further questions or contributions, feel free to open an issue or submit a pull request! ✨