Probabilities

There are four main ways of expressing the prediction from a logistic regression model – we'll look at each of them over the next four exercises. Firstly, since the response variable is either "yes" or "no", you can make a prediction of the probability of a "yes". Here, you'll calculate and visualize these probabilities.

Three variables are available:

mdl_churn_vs_relationship is the logistic regression model of has_churned versus time_since_first_purchase.
explanatory_data is a data frame of explanatory values.
plt_churn_vs_relationship is a scatter plot of has_churned versus time_since_first_purchase with a smooth glm line.

dplyr is loaded.

Use the model, mdl_churn_vs_relationship, and the explanatory data, explanatory_data, to predict the probability of churning. Assign the predictions to the has_churned column of a data frame, prediction_data. Remember to set the prediction type.