Dataset from: https://www.kaggle.com/datasets/muhammadshahidazeem/customer-churn-dataset
Models used for churn prediction (implemented from scratch):
- Linear Regression
- Logistic Regression
- K-Nearest Neighbors
Respective Accuracies for the models:
- 82.73%
- 82.84%
- 90.25%
Ideas to get a better accuracy:
- Apply both Ridge and Lasso Regularization
- Hyperparameter tuning for:
- Lambda
- Alpha
- Learning rate
Open to other suggestions!