My project is to predict what payment option a customer might use depending on various factors on a checkout screen.
For example here are some of the fields I would have
Variables : User_Location Product_Category First_Time_Purchase Repeat_Purchase Historical_Payment_Method Payment_Option_Chosen
Variable Values: China Electronics No Yes Wallet Wallet
Apologies for the bad formatting, was not sure how else to portray this. Those are some of the example fields I'd like to use and create personas of. The issue is the data is not readily available anywhere. An idea I had was to create some sort of system for generating synthetic data about transactional info.
The data has to be teach the model, lets say someone from China wants to buy fashion item, on their android, they will most likely choose x payment method, however they may also choose y payment method. There has to patterns the ML can recognise and learn from. The issue is I have no idea how to generate 100s of these personas that will allow a ML model to properly learn from these patterns. Perhaps fine tune a LLM that could generate rules based on these type of examples.
Is there a method to do this properly ?
Any suggestions or pointers would be greatly appreciated, as its been a problem I've had for a long time now.
