Research Article
Selecting Critical Data Flows in Android Applications for Abnormal Behavior Detection
Table 4
Performance comparison of precategory outlier detection on dataset
.
| Category | ā | ā | AUC of all flows | AUC of critical flows |
| LOCATION_INFORMATION | 1168 | 551 | 0.9915 | 0.9985 | NETWORK_INFORMATION | 2254 | 3068 | 0.9953 | 0.9954 | DATABASE_INFORMATION | 1525 | 815 | 0.9978 | 0.9989 | UNIQUE_IDENTIFIER | 972 | 1888 | 0.9942 | 0.9984 | NO_SENSITIVE_SOURCE | 2790 | 3883 | 0.8973 | 0.9177 | CALENDAR_INFORMATION | 1275 | 499 | 0.9399 | 0.9892 | BLUETOOTH_INFORMATION | 183 | 139 | 0.9953 | 1.0000 | ACCOUNT_INFORMATION | 320 | 33 | 0.9981 | 0.9988 |
|
|