Research Article

FraudMiner: A Novel Credit Card Fraud Detection Model Based on Frequent Itemset Mining

Algorithm 1

Training algorithm.
Input: Customer Transactions Database , support
Output: Legal Pattern Database LPD, Fraud Pattern Database FPD
Begin
                Group the transactions of each customer together.
                Let there are “” groups corresponds to “” customers
                for   to   do
                      Separate each group Gi into two different groups LGi
                            and FGi of legal and fraud transactions. Let there
                            are “” legal and “” fraud transactions
                      FIS = Apriori(LGi, , ); //Set of frequent itemset
                      LP = ; //Large Frequent Itemset
                      LPD() = LP;
                      FIS = Apriori(FGi, , ); //Set of frequent itemset
                      FP = ; //Large Frequent Itemset
                      FPD() = FP;
                endfor
                return LPD & FPD;
End