3 Machine learning for classification

In this chapter, we are going to use machine learning to predict churn.

Churn is a process in which customers stop using the services of a company. Thus, churn prediction is about identifying customers who are likely to cancel their contracts soon. If the company can do that, it can offer discounts on these services and this way keep the users.

Naturally, we can use machine learning for that: we can use the past data about customers who churned and, based on that, create a model for identifying present customers who are about to go away. This is a binary classification problem. The target variable that we want to predict is categorical and has only two possible outcomes: churn or not churn.

3.1           Churn prediction project

3.1.1   Telco churn dataset

3.1.2   Initial data preparation

3.1.3   Exploratory data analysis

3.1.4   Feature importance

3.1.5   Churn rate

3.1.6   Risk ratio

3.1.7   Mutual information

3.1.8   Correlation coefficient

3.2           Feature engineering

3.2.1   One-hot encoding for categorical variables

3.3           Machine learning for classification

3.3.1   Logistic regression

3.3.2   Training logistic regression

3.3.3   Model interpretation

3.3.4   Using the model

3.4           Next steps

3.4.1   Exercises

3.4.2   Other projects

3.5           Summary

3.6           Answers to exercises