### Table 12. Fraud Data Set, Continuous Data, Neural Network

"... In PAGE 7: ... The neural network model based on continuous data was not as good as the neural network model applied to categorical data, except that the degeneration for the training set of 120 was not as severe. Table12 shows relative accuracy for the neural network model applied to the continuous data. ... ..."

Cited by 1

### Table 9. Fraud Data Set, Categorical Data, Neural Network

"... In PAGE 6: ... For the logistic regression model, this led to a case where the test set contained 31 cases not covered by the training set. Table9 shows results for the neural network model applied to categorical data. The neural network model applied to categorical data was quite stable until the last training set where there were only 120 observations.... ..."

Cited by 1

### Table 1. Comparison of the HCMAC neural network with the MHCMAC neural network Models

"... In PAGE 15: ... D. Comparison of HCMAC Neural Network with the MHCMAC Neural Network Table1 compares the HCMAC neural network with the MHCMAC neural network in terms of memory requirement, topology structure and input feature assignment approach. Table 1 shows that the memory requirement of the original HCMAC neural network grows with the power 2 of the ceiling logarithm of the input dimensions, but the memory requirement of the MHCMAC neural network grows only linearly with the input feature dimensions.... In PAGE 15: ... Comparison of HCMAC Neural Network with the MHCMAC Neural Network Table 1 compares the HCMAC neural network with the MHCMAC neural network in terms of memory requirement, topology structure and input feature assignment approach. Table1 shows that the memory requirement of the original HCMAC neural network grows with the power 2 of the ceiling logarithm of the input dimensions, but the memory requirement of the MHCMAC neural network grows only linearly with the input feature dimensions. Moreover, the learning structure of the self-organizing HCMAC neural network is expanded based on a full binary tree topology, but the MHCMAC neural network is expanded based on an exact binary tree topology.... ..."

### Table 3. A categorization of difierent credit card fraud detection systems based on the outlier detection technique being used.

2007

Cited by 1

### Table 2 Fraud detection vs. confidence

"... In PAGE 3: ... This is also true when we use the real proportion for legal vs. misuse transactions of 1000:1 which are shown in round brackets in Table2 . Additionally, the diagnosis performance is even better than the constant, stupid diagnosis mentioned before and noted in the last table row.... ..."

### Table I. General Behavior-Based Analysis Internet Applications Application Description and Variations: Internet fraud detection Unauthorized outgoing email, unauthenticated email, unauthorized transactions Malicious email detection Spam, viruses, worms

1998

Cited by 3

### Table 1. Comparison of fraud detection probabilities and efficiencies

"... In PAGE 10: ... 7). Table1 illustrates a comparison of fraud detection probabilities and efficien- cies of the proposed scheme, A.... ..."

### Table 6 Fraud detection vs. confidence

"... In PAGE 8: ... However, when we select only those rules which also preserve their confidence sufficiently on the whole trans- action set, we obtain 510 rules. Certainly, with less rules the fraud diagnosis probability decreases slightly, but, as we see in Table6 , the confidence in the diagnosis is dra- matically increased up to 75 % due to the high proportion of legal data which are less misclassified. This is also true when we use the real proportion for legal vs.... In PAGE 8: ... This is also true when we use the real proportion for legal vs. misuse transactions of 1000:1 which are shown in round brackets in Table6 . The total diagnosis performance is even better than the constant, stupid diagnosis mentioned before and noted in the last table row.... ..."

### Table 2. Neural network modeling

2006

"... In PAGE 5: ... In model A, we calculated the IBIs from the steady-state solution, whereas in model B we continuously solved the network without discarding the transients. The model param- eters are summarized in Table2 . The parameters were the same as previously described (4), except for TNI1 (see Table 2).... ..."