I'm looking for some help understanding what the appropriate method would be for the classification problem(well really I want the probabilities) at hand.
The event that is being predicted is whether or not the customer "defaults" on a loan within the next 12 months. Observations of this are recorded every month for each customer for the last year. However, if a customer defaults then they will not appear in the dataset whilst they still remain in default (minimum 12 months), the same is true if they pay off the loan.
I would try logistic regression but I am not sure this is appropriate because of the repeated measurements.