DroidChain: A novel Android malware detection method based on behavior chains

摘要: Abstract The drastic increase of Android malware has led to strong interest in automating analysis. In this paper, fight against variants and zero-day malware, we proposed DroidChain: a method combining static analysis behavior chain model. We transform the detection problem into more accessible matrix form. Using method, propose four kinds models, including privacy leakage, SMS financial charges, installation, privilege escalation. To reduce time complexity, WxShall-extend algorithm. had moved prototype GitHub evaluate using 1260 samples. Experimental results demonstrate accuracy, precision, recall 73%–93%, 71%–99%, 42%–92%, respectively. Calculation accounts for 6.58% well-known Warshall algorithm’s expense. Results that our which can detect simultaneously, is better than Androguard Kirin.