The following project is about an ISP (Internet Service Provider) with customer data. The goal is to generate 3 ML models to predict churn, and will use 3 classification models.
Data corresponds to a fictional ISP/Telco in USA with a total of 7043 customers. There are many different files that concatenate on the Customer_id.
The models used for the classification problem are
- Logaritmic Regression
- KNN
- Decision trees
These models are later Hyperparameter Tuned.
At the end of the Notebok, there are some conclusion and Next Steps recommendations for the ISP to prevent churn.