How to predict an outcome within a specific time window?

Question

I have a dataset which has around 10K records.

My objective is to predict whether the customer will churn or not. Binary classification problem with each class representing around 55:45 proportion and 20 features.

I understand when it's just about prediction, I can apply some binary classification algorithm and find out whether the customer churns or not

But how do I incorporate the objective of finding whether the customer will churn in 30 days or not?

Another example is find whether patient will be dead within 30 days from the date of discharge. I have his date of discharge along with other features like Blood pressure, Cholesterol etc.

Rather than just predicting whether he will be dead or not anytime in future, I would like to restrict it to 30 days from date of discharge.

Hope I gave the details to help you understand the question better.

Answer 1

It depends how deep technichally you want to go. You can apply a slight modification of a Survival methods/ cox models that relate the time that passes, before some event occurs, to one or more covariates that may be associated with that quantity of time.

Also if you group de features you can make the problem look like as a classical binary classification problem. But you should do a bit of data engineering in order to have the labels this good.

Probably the easiest is to modify your data and make it look like a classification problem when the target is if the person went away in the next months.

Answer 2

Beeing X in the future and beeing X in specific time in the future is just a subset of the first one.

So what one really needs to do is just determine the probabilities (or parameters that give us these probabilities)

P(X|t>30)

Where you can model t, also as your feature. So just fit a model on this data, where you have mutliclassification of:

1. dead within 30 days
2. dead after 30 days
3. alive after 30 days