Skip to content

Latest commit

 

History

History
107 lines (72 loc) · 2.98 KB

README.md

File metadata and controls

107 lines (72 loc) · 2.98 KB

🧠 Customer Churn Prediction

This project predicts whether a customer is likely to churn from a bank using a machine learning model built with TensorFlow and deployed via Streamlit. The model is trained on customer profile data and financial metrics.

📊 Dataset

The dataset used is Churn_Modelling.csv and contains the following features:

  • CreditScore
  • Geography
  • Gender
  • Age
  • Tenure
  • Balance
  • NumOfProducts
  • HasCrCard
  • IsActiveMember
  • EstimatedSalary

The target variable is Exited, indicating whether a customer has churned.

📁 Project Structure


.
├── app.py                        # Streamlit app for churn prediction
├── streamlit\_regression.py      # Streamlit app for salary regression (optional)
├── model.h5                     # Trained churn prediction model
├── regression\_model.h5          # Trained regression model
├── scaler.pkl                   # StandardScaler for numerical features
├── lable\_encoder\_gender.pkl     # LabelEncoder for Gender
├── one\_hot\_encoder\_geo.pkl      # OneHotEncoder for Geography
├── requirements.txt             # Python dependencies
├── runtime.txt                  # Runtime environment (for deployment)
├── hyperparameter\_tuning.ipynb  # Notebook for tuning hyperparameters
├── predictions.ipynb            # Model prediction and evaluation
├── salaryregression.ipynb       # Regression model training
├── experiments.ipynb            # Initial exploratory analysis
├── Churn\_Modelling.csv          # Dataset
├── README.md                    # Project documentation

🚀 How to Run

  1. Install dependencies:
pip install -r requirements.txt
  1. Run the Streamlit app:
streamlit run app.py

The app will launch in your browser where you can input customer details and see churn predictions.

🧩 Model Workflow

  1. Preprocessing

    • Encode categorical features:

      • Gender → LabelEncoded
      • Geography → OneHotEncoded

    • Scale numerical features with StandardScaler.

  2. Model

    • A deep learning classification model built using Keras.
    • Output is a churn probability (between 0 and 1).

  3. Prediction

    • If the probability > 0.5 → Customer is likely to churn.
    • Else → Customer is likely to stay.

📈 Experimentation

  • Hyperparameter tuning notebooks are available for training custom models.

  • You can find:

    • hyperparameter_tuning.ipynb — tuning classification model
    • salaryregression.ipynb — regression analysis
    • predictions.ipynb — model inference and evaluation

🛠 Deployment

  • requirements.txt and runtime.txt are included for deployment on platforms like Heroku or Streamlit Cloud.
  • For containerized deployment, a Dockerfile can be added if needed.