The objective of the churn prediction model is to predict the customers that are likely to stop using PayTV service hrough the usage behavior.
Data preparation for churn prediction starts with aggregating all available information about the customer. The data that is obtained for predicting the churn is classified in the following categories:
- PayTV Contract data such as life of contract, region, profile box type, payment data
- Duration Behavior data such as duration by day, month, channel,..
- Internet usage data such as download/upload size
- Demographic data
Folders:
-
member_folder
- hieunt124
- lanhnv3
-
production
- conf
- main_config.py
- utils_con.py
- service
- data
- main
- build_features.py
- train_model.py
- predict_model.py
- conf