Methods and apparatus for evaluating classification performance of risk engine models
Methods and apparatus are provided for evaluating the classification performance of different risk engine models. A classification performance of an authentication method is evaluated by obtaining performance data for an authentication method; generating a receiver operating characteristic (ROC) curve for the obtained performance data; determining a partial area under the curve (pAUC) for a region of interest of the ROC curve; and providing a performance score for the authentication method based on the pAUC. The region of interest comprises, for example, a region of false positives. The pAUC is optionally standardized using a McClish Transformation. The performance score for the authentication method can be compared to a second performance score for a second authentication method. A confidence level can optionally be provided for the comparison based on a natural test statistic.
1. A method, comprising:
obtaining performance data for a first authentication method;
generating, using at least one processing device, a receiver operating characteristic (ROC) curve for said obtained performance data;
determining a partial area under the curve (pAUC) for a region of interest of the ROC curve;
providing a first performance score for the first authentication method based on the pAUC;
storing said first performance score in at least one record of at least one memory;
comparing, using said at least one processing device, said first performance score for said first authentication method to a second performance score for a second authentication method; and
selecting, using said at least one processing device, one of said first authentication method and said second authentication method for authenticating one or more users based on the comparison and a confidence level for a result of said comparison of said first performance score for said first authentication method and said second performance score for said second authentication method based on a natural test statistic that is a function of the pAUC.
2. The method of claim 1 , wherein said region of interest comprises a region of false positives.
3. The method of claim 1 , further comprising the step of standardizing the pAUC using a McClish Transformation.
4. The method of claim 1 , wherein said natural test statistic is based on McClish Indices η 1 and η 2 for said first authentication method and said second authentication method, respectively.
5. The method of claim 1 , wherein said ROC curve comprises a true-positive-rate as a function of a false-positive-rate.
6. The method of claim 1 , wherein said pAUC evaluates a portion of the ROC curve.
7. The method of claim 1 , wherein said first performance score characterizes a classification performance of said first authentication method.
8. The method of claim 1 , wherein said first authentication method comprises an evaluation of one or more of a password, a one-time passcode, biometric information, a token and a certificate.
9. The method of claim 1 , wherein said first authentication method comprises a current authentication method and said second authentication method comprises an updated authentication method.
10. A non-transitory machine-readable storage medium having encoded therein executable code of one or more software programs, wherein the one or more software programs when executed by at least one processing device perform the following steps:
obtaining performance data for a first authentication method;
generating a receiver operating characteristic (ROC) curve for said obtained performance data;
determining a partial area under the curve (pAUC) for a region of interest of the ROC curve;
providing a first performance score for the first authentication method based on the pAUC;
storing said first performance score in at least one record of at least one memory;
comparing, using said at least one processing device, said first performance score for said first authentication method to a second performance score for a second authentication method; and
selecting, using said at least one processing device, one of said first authentication method and said second authentication method for authenticating one or more users based on the comparison and a confidence level for a result of said comparison of said first performance score for said first authentication method and said second performance score for said second authentication method based on a natural test statistic that is a function of the pAUC.
11. The non-transitory machine-readable storage medium of claim 10 , wherein said first performance score characterizes a classification performance of said first authentication method.
12. An apparatus, the apparatus comprising:
a memory; and
at least one processing device, coupled to the memory, operative to:
obtain performance data for a first authentication method;
generate, using said at least one processing device, a receiver operating characteristic (ROC) curve for said obtained performance data;
determine a partial area under the curve (pAUC) for a region of interest of the ROC curve;
provide a first performance score for the first authentication method based on the pAUC;
store said first performance score in at least one record of at least one memory;
compare, using said at least one processing device, said first performance score for said first authentication method to a second performance score for a second authentication method; and
select, using said at least one processing device, one of said first authentication method and said second authentication method for authenticating one or more users based on the comparison and a confidence level for a result of said comparison of said first performance score for said first authentication method and said second performance score for said second authentication method based on a natural test statistic that is a function of the pAUC.
13. The apparatus of claim 12 , wherein said region of interest comprises a region of false positives.
14. The apparatus of claim 12 , wherein said at least one processing device is further configured to standardize the pAUC using a McClish Transformation.
15. The apparatus of claim 14 , wherein said natural test statistic is based on McClish Indices η 1 and η 2 for said first authentication method and said second authentication method, respectively.
16. The apparatus of claim 12 , wherein said ROC curve comprises a true-positive-rate as a function of a false-positive-rate.
17. The apparatus of claim 12 , wherein said pAUC evaluates a portion of said ROC curve.
18. The apparatus of claim 12 , wherein said apparatus is implemented as part of a processing platform.
19. The apparatus of claim 12 , wherein said first performance score characterizes a classification performance of said first authentication method.
20. The apparatus of claim 12 , wherein said first authentication method comprises a current authentication method and said second authentication method comprises an updated authentication method.