Table 2 Computational measures calculated for the four methods with the three sets of negative examples with different sizes n1, n2, n3.
Measure Method | AUC | Likelihood Log Score | Likelihood Lift | Likelihood RMSE |
|---|---|---|---|---|
n1 = 221 | ||||
 MRAM | 0.846 | −0.259 ± 0.021 | 0.433 ± 0.020 | 0.211 ± 0.001 |
 Decision Tree | 0.837 | −0.405 ± 0.063 | 0.287 ± 0.063 | 0.213 ± 0.020 |
 Bayesian Network | 0.822 | −0.540 ± 0.175 | 0.152 ± 0.175 | 0.284 ± 0.029 |
 Neural Network | 0.783 | −0.416 ± 0.084 | 0.276 ± 0.084 | 0.224 ± 0.026 |
n2 = 500 | ||||
 MRAM | 0.890 | −0.149 ± 0.010 | 0.469 ± 0.019 | 0.113 ± 0.016 |
 Decision Tree | 0.827 | −0.294 ± 0.102 | 0.314 ± 0.102 | 0.115 ± 0.009 |
 Bayesian Network | 0.814 | −0.348 ± 0.073 | 0.276 ± 0.073 | 0.116 ± 0.031 |
 Neural Network | 0.783 | −0.501 ± 0.108 | 0.114 ± 0.108 | 0.237 ± 0.018 |
n3  =  1,000 | ||||
 MRAM | 0.883 | −0.211 ± 0.056 | 0.256 ± 0.054 | 0.063 ± 0.007 |
 Decision Tree | 0.866 | −0.265 ± 0.039 | 0.202 ± 0.040 | 0.166 ± 0.001 |
 Bayesian Network | 0.808 | −0.263 ± 0.048 | 0.218 ± 0.047 | 0.130 ± 0.025 |
 Neural Network | 0.804 | −0.293 ± 0.060 | 0.198 ± 0.057 | 0.174 ± 0.019 |
