Yep. OP's only real recourse is to just try again in 6 months or a year or whatever and hope that their ML algorithm evaluates their data differently.

If Ebay gave a credit report-style summary saying "you're banned because you're associated with this IP range" or something, then indeed this becomes information that would be exploited by fraudsters. If OP is actually innocent then their being banned is considered an acceptable risk.... one can only hope that in future model training though that this ban would be considered a false positive.