Skip to content

Latest commit

 

History

History
103 lines (59 loc) · 3.46 KB

README.md

File metadata and controls

103 lines (59 loc) · 3.46 KB
Raw

License: MIT jupyter

Overview

This project aims to predict customer churn for a telecom company using various machine learning models. Customer churn occurs when customers stop doing business with a company, and predicting churn helps companies take proactive measures to retain customers.

Dataset

The dataset contains customer information such as demographic details, services subscribed, billing information, and whether the customer churned or not. Key columns include:

  • customerID: Unique identifier for each customer.
  • gender: Gender of the customer.
  • SeniorCitizen: Whether the customer is a senior citizen (1 for Yes, 0 for No).
  • Partner: Whether the customer has a partner.
  • Dependents: Whether the customer has dependents.
  • tenure: Number of months the customer has stayed with the company.
  • PhoneService: Whether the customer has a phone service.
  • InternetService: Type of internet service (DSL, Fiber optic, No).
  • Churn: Target variable indicating if the customer churned (Yes/No).

Project Structure

  • data/: Directory containing the dataset (customer_churn.csv).
  • notebooks/: Jupyter notebooks with data exploration, preprocessing, and model training.
  • models/: Directory to save trained models.

Installation

To get started with this project, follow these steps:

  1. Clone the repository:
git clone (https://github.com/10-kp/cust_churn_project)
cd customer-churn-prediction
  1. Install the required packages:
pip install -r requirements.txt
  1. Run the Jupyter notebook:
jupyter notebook notebooks/Customer_Churn_Analysis.ipynb

Usage

  1. Data Preprocessing:

    • Handle missing values.
    • Convert categorical variables into numeric form using Label Encoding.
    • Check for multicollinearity using Variance Inflation Factor (VIF) and remove highly correlated features.

  2. Model Building:

    • Split the dataset into training and testing sets.
    • Train and evaluate Logistic Regression, Decision Tree, and Random Forest models.
    • Measure model performance using accuracy scores.

  3. Feature Engineering:

    • Experiment with feature selection and engineering to improve model performance.

  4. Evaluation:

    • Evaluate the models using accuracy, confusion matrix, and other relevant metrics.
    • Compare the performance of different models.

Results

  • Logistic Regression: Achieved an accuracy of 76.7%.
  • Decision Tree: Achieved an accuracy of 73.0%.
  • Random Forest: Achieved an accuracy of 76.8%.

The Random Forest model performed the best, making it a strong candidate for predicting customer churn.

Future Work

  • Experiment with more advanced models like XGBoost.
  • Perform hyperparameter tuning to further improve model performance.
  • Analyze feature importance to understand which features contribute most to churn prediction.

Contributing

Contributions are welcome! If you have suggestions or improvements, feel free to create a pull request or open an issue.

License

This project is licensed under the MIT License. See the LICENSE file for more details.

Acknowledgements