0
$\begingroup$

I'm using predict_survival_function method in lifelines package and I want to have the conditional survival function of each individual in my dataset. The dataset includes some users feature, the weeks from signup to churn and whether they churn or not. The result is a matrix which gives the survival probabilities of users over time. I'm a bit confused with the indices of this matrix.

I'm seeing that they start from the minimum signup_to_churn duration and goes to maximum of that. Below is an example:

    ----+-----------------+
   index| user_1 | user_2 |
    ----+-----------------+
    10.5| 0.89   | 0.87   |
    ----+-----------------+

Does that mean 10.5 weeks after signup, user_1 has 0.89 probability of survival? or it means 10.5 weeks after the date of the analysis, user_1 has 0.89 probability of survival?

Basically, given that this is a conditional probability, what are the start time of these indices? signup? or date of the analysis?

$\endgroup$
1
$\begingroup$

The index represents the duration of the subject from their signup. So in your example, after 10.5 weeks since user_1 signed up, they have a 89% chance of being alive still.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.