Optimal antimalware signatures database

1. A method for creating an exportable digital collection of optimized malware family-specific signatures for protecting a computing system from malware, the method comprising:

providing a sample of files comprising known clean files and known malware files;

wherein the known clean files and malware files comprise a plurality of attribute vectors;

identifying signatures for the plurality of attribute vectors;

calculating, for each signature from among the signatures, a false positive rate by identifying the number of clean files with each said signature, and grouping malware files that share a signature in a family cluster;

from all the signatures in the family cluster, selecting a signature with the lowest false positive rate as a representative signature for the family cluster, wherein the representative signature is a first signature and a second representative signature is selected and wherein the second representative signature is selected using a predefined target function, wherein the predefined target function selects, as the second representative signature, the signature with the greatest probability to be found among all samples of the malware family cluster; and

exporting the representative signature configured for use by a scan engine.

2. The method according to claim 1 , wherein the predefined target function selects, as a second representative signature, the signature with the second lowest false positive rate and the greatest probability to be found among all samples of the malware family cluster.

3. The method according to claim 2 , wherein the predefined target function further selects, as a third representative signature, the signature that requires the least amount of resources to scan a file for the signature.

4. The method according to claim 1 , wherein the predefined target function further selects, as a second representative signature, the signature that requires the least amount of resources to scan a file for the signature.

5. The method according to claim 1 , wherein the predefined target function further selects as a third representative signature, the signature that requires the least amount of resources to scan a file for the signature.

6. The method according to claim 1 , wherein the representative signature is further selected by a predefined target function that rejects the signature with the lowest false positive rate when the average amount of resources required to scan a file for the signature exceeds a predetermined threshold.

7. The method according to claim 1 , wherein the attribute vectors include at least one of binary n-grams, strings, or opcodes.

8. The method according to claim 1 , wherein the sample further comprises statistics of all possible strings or binary sequences for the files in the sample.

9. A system for creating a collection with optimized family-specific signatures for protecting from malware, the system comprising:

a processor coupled to a memory storing instructions;

wherein the processor is configured to analyze a sample of clean and malware files, collect statistics of potential signatures for the clean and malware files as subsets of chosen attribute vectors,

and calculate, for each potential signature, a false positive rate by identifying the number of clean files with that potential signature;

wherein the processor is further configured to group malware files that share a signature in a family cluster, and from all the potential signatures, to choose a representative set of signatures for the malware file family based on a predefined target function, wherein from among the representative set of signatures, a representative signature with the lowest false positive rate is selected from the family cluster as a first signature and a second representative signature is selected and wherein the second representative signature is selected using a predefined target function, wherein the predefined target function selects, as the second representative signature, the signature with the greatest probability to be found among all samples of the malware family cluster; and

wherein the processor is further configured to export a version of the collection with optimized malware family-specific signatures configured to be implemented by a scan engine.

10. The system according to claim 9 , wherein the predefined target function is configured to calculate a probability to find a signature in the clean files.

11. The system according to claim 10 , wherein the predefined target function is further configured to calculate a probability to find a signature among all samples for a file of the malware family files.

12. The system according to claim 11 , wherein the predefined target function is further to calculate an average amount of resources required to scan a file for the signature.

13. The system according to claim 10 , wherein the predefined target function is further configured to calculate an average amount of resources required to scan a file for the signature.

14. The system according to claim 9 , wherein the predefined target function is further configured to calculate a probability to find a signature among all samples for a file of the malware family files.

15. The system according to claim 14 , wherein the predefined target function is further configured to calculate an average amount of resources required to scan a file for a signature.

16. The system according to claim 9 , wherein the predefined target function is further configured to calculate an average amount of resources required to scan a file for a signature.

17. The system according to claim 9 , wherein the attribute vectors include at least one of binary n-grams, strings, or opcodes.

18. The system according to claim 9 , wherein the sample further comprises statistics of all possible strings or binary sequences for the files in the sample.

19. The method according to claim 1 , wherein the predefined target function incorporates at least one of stability of the signature, generalizability of the signature, and overlap of the signature with clean files.

20. The system according to claim 9 , wherein the predefined target function incorporates at least one of stability of the signature, generalizability of the signature, and overlap of the signature with clean files.