Customer Personality Analysis and Churn

• This is a quickly whipped up, well structured project using a Customer Personality dataset.
• I have conducted a quite in-depth feature extraction (as outlined in feature_extraction.ipynb).
• Models were tinkered with in train.ipynb.
• Execute main_train.py using python main_train.py.
• Currently implemented models, (switch between them using the cli option --model as one of [ logistic, random_forest ]):
  1. Logistic Regression
  2. Random Forest Classifiers

Required Libraries

1. sklearn
2. matplotlib
3. scipy
4. pandas
5. numpy
6. seaborn

Model metrics:

Logistic Regression

1. Confusion Matrix -
   
   

2. ROC Curve -
   
   

3. Test Accuracy - $0.775$

Random Forest

1. Confusion Matrix -
   
   

2. ROC Curve -
   
   

3. Test Accuracy - $0.99$