Detection Technology of Mobile Application Third-Party Library Based on Binary Tree Model

At present, most of the research on the detection effect of third-party libraries is not ideal, and it has weak robustness to the third-party libraries with confusing or flat packages, and few technologies can be detected at the library version level. Aiming at the current problems, this paper proposes a third-party library detection model based on binary tree, which can effectively enhance the robustness of detection by using keywords, comments and constant strings in source code files as data sets. In addition, a library version identification technology based on maximum feature matching is proposed to identify the versions of third-party libraries. Experiments are conducted on 200 Android applications from the application market to detect the okhttp3 library and identify the version of the library. The results show that the detection accuracy of the third-party library reaches 96.5%, and the version identification accuracy of the third-party library reaches 76.84%.