Systems and methods for detecting anomalous system command line data

1. A computer-implemented method for detecting anomalous system command line data, at least a portion of the method being performed by one or more computing devices comprising at least one processor, the method comprising:

receiving, by the one or more computing devices, command line data from a target computing system;

building, by the one or more computing devices, a baseline model that utilizes machine-learning to analyze the command line data, the baseline model comprising a support-vector machine (SVM), natural language processing, and a hashing function;

assigning, by the one or more computing devices and utilizing the baseline model, a score to each of a plurality of instances of the command line data; and

identifying, by the one or more computing devices and based on the score, anomalous commands comprising potentially malicious data when any of the instances of the command line data fails to exceed a threshold.

2. The computer-implemented method of claim 1 , further comprising performing a security action that includes reporting the anomalous commands comprising the potentially malicious data to a network security administrator for analysis.

3. The computer-implemented method of claim 1 , wherein building the baseline model comprises:

extracting, utilizing the natural language processing, lexical information comprising clean data from the command line data;

generating a one-class SVM from the clean data; and

combining the SVM with the hashing function to build the baseline model.

4. The computer-implemented method of claim 1 , wherein assigning the score comprises:

receiving the command line data instances in the hashing function;

identifying a target data instance in the command line data instances having commands that are present in a clean index of the hashing function; and

assigning a clean data score to the target data instance.

5. The computer-implemented method of claim 1 , wherein assigning the score comprises:

receiving the command line data instances in the hashing function;

identifying a target data instance in the command line data instances having commands that are absent from a clean index in the hashing function;

determining that the commands in the target data instance exceed the threshold; and

assigning a clean data score to the target data instance.

6. The computer-implemented method of claim 1 , wherein assigning the score comprises:

receiving the command line data instances in the hashing function;

identifying a target data instance in the command line data instances having commands that are absent from a clean index in the hashing function;

determining that the commands in the target data instance fail to exceed the threshold;

receiving the target data instance in the SVM from the hashing function;

determining, by the SVM, that the commands in the target data instance are clean; and

assigning a clean data score to the target data instance.

7. The computer-implemented method of claim 6 , wherein determining, by the SVM, that the target data instance is clean comprises determining that the commands comprise at least one of previously seen commands or a variant of previously seen commands in a clean index of the hashing function.

8. The computer-implemented method of claim 6 , wherein the commands comprise near duplicate commands utilized by a dual-use application on the target computing system.

9. The computer-implemented method of claim 6 , wherein the commands comprise low reputation data.

10. The computer-implemented method of claim 1 , wherein assigning the score comprises:

receiving the command line data instances in the hashing function;

identifying a target data instance in the command line data instances having commands that are absent from a clean index in the hashing function;

determining that the commands in the target data instance fail to exceed the threshold;

receiving the target data instance in the SVM from the hashing function;

determining, by the SVM, that the commands in the target data instance are uncommon commands; and

assigning an anomalous data score to the target data instance.

11. The computer-implemented method of claim 1 , wherein the threshold comprises command line data samples associated with high reputation data.

12. The computer-implemented method of claim 1 , wherein the hashing function comprises a minhash function and a locality-sensitive hashing index.

13. A system for detecting anomalous system command line data, the system comprising:

at least one physical processor;

physical memory comprising computer-executable instructions and one or more modules that, when executed by the physical processor, cause the physical processor to:

receive, by a receiving module, command line data from a target computing system;

build, by a machine-learning module, a baseline model that utilizes machine-learning to analyze the command line data, the baseline model comprising a support-vector machine (SVM), natural language processing, and a hashing function;

assign, by a scoring module and utilizing the baseline model, a score to each of a plurality of instances of the command line data; and

identify, by an identification module and based on the score, anomalous commands comprising potentially malicious data when any of the instances of the command line data fails to exceed a threshold.

14. The system of claim 13 , wherein the machine-learning module builds the baseline model by:

extracting, utilizing the natural language processing, lexical information comprising clean data from the command line data;

generating a one-class SVM from the clean data; and

combining the SVM with the hashing function to build the baseline model.

15. The system of claim 13 , wherein the scoring module assigns the score by:

receiving the command line data instances in the hashing function;

identifying a target data instance in the command line data instances having commands that are present in a clean index of the hashing function; and

assigning a clean data score to the target data instance.

16. The system of claim 13 , wherein the scoring module assigns the score by:

receiving the command line data instances in the hashing function;

identifying a target data instance in the command line data instances having commands that are absent from a clean index in the hashing function;

determining that the commands in the target data instance exceed the threshold; and

assigning a clean data score to the target data instance.

17. The system of claim 13 , wherein the scoring module assigns the score by:

receiving the command line data instances in the hashing function;

identifying a target data instance in the command line data instances having commands that are absent from a clean index in the hashing function;

determining that the commands in the target data instance fail to exceed the threshold;

receiving the target data instance in the SVM from the hashing function;

determining, by the SVM, that the commands in the target data instance are clean; and

assigning a clean data score to the target data instance.

18. The system of claim 17 , wherein the SVM determines that the target data instance is clean by determining that the commands comprise at least one of previously seen commands or a variant of previously seen commands in a clean index of the hashing function.

19. The system of claim 17 , wherein the commands comprise near duplicate commands utilized by a dual-use application on the target computing system.

20. A non-transitory computer-readable medium comprising one or more computer-executable instructions that, when executed by at least one processor of a computing device, cause the computing device to:

receive command line data from a target computing system;

build a baseline model that utilizes machine-learning to analyze the command line data, the baseline model comprising a support-vector machine (SVM), natural language processing, and a hashing function;

assign, utilizing the baseline model, a score to each of a plurality of instances of the command line data; and

identify, based on the score, anomalous commands comprising potentially malicious data when any of the instances of the command line data fails to exceed a threshold.