Machine learning-based malware detection system and method
Disclosed is a computer implemented method for malware detection that analyses a file on a per packet basis. The method receives a packet of one or more packets associated a file, and converting a binary content associated with the packet into a digital representation and tokenizing plain text content associated with the packet. The method extracts one or more n-gram features, an entropy feature, and a domain feature from the converted content of the packet and applies a trained machine learning model to the one or more features extracted from the packet. The output of the machine learning method is a probability of maliciousness associated with the received packet. If the probability of maliciousness is above a threshold value, the method determines that the file associated with the received packet is malicious.
1. A computer program product embodied in a non-transitory computer readable storage medium and comprising computer instructions that when executed by a processor cause the processor to perform steps of:
receiving a first packet of one or more packets associated with a file;
determining a file type of the file from the first packet;
converting contents of the first packet into a corresponding digital representation for feature extraction;
extracting one or more features from the corresponding digital representation; and
applying a trained machine learning model to the one or more features to determine probability of maliciousness.
2. The computer program product embodied in a non-transitory computer readable storage medium of claim 1 , wherein the steps further include
labeling the file based on the probability of maliciousness.
3. The computer program product embodied in a non-transitory computer readable storage medium of claim 1 , wherein the steps further include
responsive to the first packet having the probability of maliciousness as benign, receiving a next packet of the one or more packets; and
performing the converting, extracting, and applying on the next packet.
4. The computer program product embodied in a non-transitory computer readable storage medium of claim 3 , wherein the steps further include
continuing with additional packets of the one or more packets until a determination of whether the file is malicious or benign.
5. The computer program product embodied in a non-transitory computer readable storage medium of claim 1 , wherein the file type is one of a portable executable (PE) file, a portable document format (PDF) file, a Dynamic Loaded Library (DLL), a JavaScript (JS) file, a Hypertext Markup Language (HTML) file, and a Microsoft Office File.
6. The computer program product embodied in a non-transitory computer readable storage medium of claim 1 , wherein the digital representation is any of a decimal representation, a binary representation, a hexadecimal representation, a tokenized script, and a tokenized domain.
7. The computer program product embodied in a non-transitory computer readable storage medium of claim 1 , wherein the trained machine learning model comprises one or more decision trees.
8. The computer program product embodied in a non-transitory computer readable storage medium of claim 1 , wherein the one or more features include n-gram features.
9. The computer program product embodied in a non-transitory computer readable storage medium of claim 1 , wherein the one or more features include an entropy feature.
10. The computer program product embodied in a non-transitory computer readable storage medium of claim 1 , wherein the one or more features include a domain feature.
11. A method comprising steps of:
receiving a first packet of one or more packets associated with a file;
determining a file type of the file from the first packet;
converting contents of the first packet into a corresponding digital representation for feature extraction;
extracting one or more features from the corresponding digital representation; and
applying a trained machine learning model to the one or more features to determine probability of maliciousness.
12. The method of claim 11 , wherein the steps further include
labeling the file based on the probability of maliciousness.
13. The method of claim 11 , wherein the steps further include
responsive to the first packet having the probability of maliciousness as benign, receiving a next packet of the one or more packets; and
performing the converting, extracting, and applying on the next packet.
14. The method of claim 13 , wherein the steps further include
continuing with additional packets of the one or more packets until a determination of whether the file is malicious or benign.
15. The method of claim 11 , wherein the file type is one of a portable executable (PE) file, a portable document format (PDF) file, a Dynamic Loaded Library (DLL), a JavaScript (JS) file, a Hypertext Markup Language (HTML) file, and a Microsoft Office File.
16. The method of claim 11 , wherein the digital representation is any of a decimal representation, a binary representation, a hexadecimal representation, a tokenized script, and a tokenized domain.
17. The method of claim 11 , wherein the trained machine learning model comprises one or more decision trees.
18. The method of claim 11 , wherein the one or more features include n-gram features.
19. The method of claim 11 , wherein the one or more features include an entropy feature.
20. The method of claim 11 , wherein the one or more features include a domain feature.