UserChurnRate_supervised-learning

This is a project to build supervised ML classification models to predict users' churn rate

There are seven sections in the project:

1.Data loading, observation(distribution, correlation, stripping)

2.Identify classification/regression problem, consider possible models

3.Feature Engineering, One-hot-encode (Linearty), drop (unrelated, univariance, low variance), split out ground truth

4.Model Selection, CV

5.Hyperparameter tuning, GridSearchCV

6.Model Evaluation, Confusion Matrix, ROC, AUC, MSE

7.Feature Selection, Coefficient, optimal number of features, RF feature importance

Probablity option for classifiers

loop through 3,4,5,6,7 for better model and better solution

source data attached Please use Jupyter Notebook to open