关键词: 便携式文档格式, 恶意代码检测, 频谱变换, 机器学习, 特征提取

Abstract: In basic research based on compound documents malicious code detection method of spectrum analysis, a spectrum analysis method based on machine learning is proposed to detect the malicious code in PDF file, firstly, we introduced the PDF file malicious code detection system and make a further optimization of the improved real sequence FFT algorithm based on the original ; we made the spectrum transformation of all the field contents extracted from the PDF file storage structure, and analyzed the transformed spectrum map, extracted feature attributes of those spectrum maps by using machine learning algorithms, then formed whether PDF files infected with malicious code is detected according to the machine learning algorithm ,and finally, through experimental analysis , to verify the correctness and validity of the method used , and gives the implementation of the whole system and test result data.

Key words: portable document format (PDF), malicious code detection, spectrum transform, machine learning, feature extraction