Open AccessMethod for constructing multi-dimensional feature map of malicious code
Author(s) -
Haocong Ma,
Zhang Ji,
Junhua Zhou,
Xiang Zhai,
Junjie Xue,
Hang Ji
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1748/4/042055
Subject(s) - computer science , code (set theory) , construct (python library) , feature (linguistics) , graph , convolutional neural network , data mining , theoretical computer science , artificial intelligence , programming language , philosophy , linguistics , set (abstract data type)
Malicious code is characterized by a large number of types, rapid increase in number, continuous update of transmission routes, and continuous enhancement of back analysis and back detection methods. Therefore, how to effectively detect and analyze malicious code has been a problem of great concern. This paper studies the features of binary file and disassembly file of malicious code, introduces the concept of information gain, and proposes a method to construct the multi-dimensional characteristic graph of malicious code. Finally, the convolutional neural network is used to classify the multi-dimensional feature graph of malicious code, which provides a new idea for the feature extraction of malicious code.