The way I would approach this is to get your Fraud team to supply a list of card numbers recently hit with CPP fraud. Then I would analyse the transactions for those cards over the same period to see if I could find a common behaviour pattern linked to CPP for these cards versus 'normal' transaction behaviour.
I would then develop and test an algorithm against a population of credit card transactions to measure how effective it is in identifying CPP. This is not a trivial task and will involve processing millions of card transactions and many weeks of effort with a lot of trial and error to come up with an effective solution.
Note: every credit card providers' transaction data is different so it is highly-unlikely an off-the-shelf solution will work without a lot of customisation. DIY (Do It Yourself) is probably a better bet.