Overview
In a group of 6, we used a telecom company’s customers’ database to predict their customer churn using several machine learning models.
We have a database with demographic and account-related information about the customers and the services they subscribed to in the telecommunication company.
First, we cleaned and transformed this dataset. We explored our data by visualizing and understanding it to know which information is more useful for our analysis and if some adjustments are needed. Then we used some prediction models:
- Logistic regression.
- Random forests.
- Support vector machines.
We tested these models and evaluated their performance in predicting customer churn.
Finally, we concluded by selecting the best model with hyperparameters.
We focused our report and our analysis on those questions:
- Identify the customers who are going to churn.
- Features that correlate to customer churn.
- Different types of churns.
- Recommendations based on the problems discovered.