Telco Customer Churn Prediction

Objectives: Build predictive model to predict customer churn

Models:

  • KNN
  • Logistic Regression
  • SVM
  • Decision Tree
  • DNN
  • Customer churn, also known as customer attrition, is the loss of clients or customers by a business. Telephone or Internet service, insurance companies, and other services based on subscription (e.g. Netflix, Spotify, etc.) Customer churn rate is often used as a key business metric for these types of companies because the cost of acquiring a new customer is far more than retaining an existing one. By applying a predictive model, the company could take the initiative to prevent losing customers.

    Result from my medels:

    Model Cross-validation Score Auc Score
    KNN 0.83 0.71
    Logistic Regression 0.84 0.71
    SVC 0.84 0.72
    SVM 0.83 0.71
    Decision Tree 0.82 0.72

    Result after applying PCA:

    Model Auc Score Auc Score_PCA
    KNN 0.71 0.69
    Logistic Regression 0.71 0.68
    SVC 0.72 0.67
    SVM 0.71 0.69
    Decision Tree 0.72 0.69


    (Please find Python code here.)