Structural Feature Based Anomaly Detection for Packed Executable Identification - Computational Intelligence in Security for Information Systems - page 279

Total Interruption Time (TIT) = JCIT + SRIT + LPIT. (3)
Unobtrusiveness(U)=(TIT / TTW) * 100. (4)
Average Unobtrusiveness (AU)=(U Best +U Medium+U Worst)/3(for each ex-
periment). (5)
The last line of the table 3 shows Average Unobtrusiveness of the Model
(AUM), that is, the result of adding all experiments unobtrusiveness divided by
the number of all cases. The experiment result shows that the AUM is 7.49%.
The factors that decrease the model unobtrusiveness are as follow:
1. The integrated privacy design of the model which considers the user privacy
from user authentication phase until saving data phase.
2. As it was described earlier, the privacy alarms only rise for APL in the range
of 25-75 which reduces the probability of rising privacy alarm.
3.2 Measuring the Model Expressiveness of Privacy Policies
The Expressiveness of Privacy Policies is determined by the support that privacy
policies and privacy mechanisms of the model provide for mandatory and discre-
tionary rules, context sensitivity, uncertainty handling, and conflict resolution.
To evaluate the model privacy policys ability to support mandatory and dis-
cretionary rules, the privacy policies should be expressed in a way that commu-
nicating parties can reliably distinguish between mandatory and discretionary
rules and the model should have some trusted mechanisms to guarantee the use
of mandatory rules on all parties. The model should provide some mechanisms
to let communicating parties to choose among discretionary rules as well.
The model privacy policies are context sensitive if these privacy policies show
context captured information, and have the ability to reflect context preferred
and current privacy levels.
The model handles uncertain situations and resolve conflicts if its privacy
policies have the ability to define these uncertain situations and the model has
suitable mechanisms to manage each one.
We implemented three different tasks with different privacy levels and to show
that the privacy files support mandatory and discretionary rules, reflect context
sensitivity, handle uncertain situations, and resolve conflicts.
The following describes how the UPM fulfills each of the above mentioned
factors:
1. Support for mandatory rules: The UPM has the ability to support manda-
tory rules through the mechanism that is used for applying owner-provided
content privacy preferences on all parties. As it was said earlier, the content
privacy preferences are mandatory rules defined by the owner and they have
to be applied on all parties, the owner provides the service contents just
after receiving the confirmation of applying content privacy policies on all
parties. The same mechanism can be applied for all other mandatory rules
as well. The mandatory rules position on each party defined in the model
are designed to be distinguishable from discretionary rules.