基于人工免疫结合余弦相似度的病毒特征提取算法  

VIRUS FEATURE EXTRACTION ALGORITHM BASED ON THE COSINE SIMILARITY OF ARTIFICIAL IMMUNE SYSTEM

作　　者：杨应华 夏勇[2] 

机构地区：[1]兰州财经大学信息中心,甘肃兰州730020 [2]西北工业大学计算机学院,陕西西安710129

出　　处：《计算机应用与软件》2017年第8期301-306,共6页Computer Applications and Software

基　　金：国家自然科学基金项目(61471297)

摘　　要：针对现有计算机病毒特征提取算法无法有效提取未知病毒和变种病毒的特征,本文借鉴人工免疫思想,提出一种基于人工免疫结合余弦相似度的病毒特征提取算法。在代码层,算法采用TF-IDF对病毒DNA进行趋向性提取建立病毒候选基因库;在基因层,算法利用可变r匹配规则提取病毒候选基因库生产病毒检测基因库;在程序层,算法采用余弦相似度算法评估待测程序与病毒的相似度,对待测程序进行识别。经仿真实验,本算法与其他病毒特征提取算法相比,在较低虚警率的情况下有较高的病毒识别率。The existing computer virus feature extraction algorithm cannot ef fect ively extract unknown viruses and variants of the virus characteristics, thus a virus feature extraction algorithm based on the cosine similarity of artificial immune system is proposed with the help of artificial immune theory. Establishing the virus candidate gene bank by adopting TF-IDF to carry on the tendency extraction to the virus DNA in the code layer algorithm; using variable r matching rule to extract virus candidate gene pool to produce virus detection gene bank in gene layer algorithm. Finally, the program layer uses the cosine similarity algorithm to evaluate the similarity between the test program and the virus, and to recognize the test program. Simulation experiments show that the proposed algorithm has higher virus recognition rate compared with other virus feature extraction algorithms in the case of low false alarm rate.

关 键 词：人工免疫 特征提取 TF-IDF算法 可变r匹配 余弦相似度 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]