Abstract:

Intent test is an important part before the release of Android applications, when the test case coverage is incomplete, the potential risk will stay in the application. This paper proposes a self-learning Intent filtering reinforcement scheme based on decision tree to extract filtering rules for the potential risks, which caused by the application without comprehensive and effective security verification of Intent communication. There is no need to modify the source or installation package, just to place the application in a safe container designed in this article. The scheme uses the decision tree algorithm to intercept the Intent attack with high similarity, and protect the application of the runtime from malicious Intent. At the same time, the algorithm has the ability of self-learning, according to the running state of current application, it can construct decision tree and generate filtering rules to adapt to the new environmental changes. The experimental results show that the reinforcement scheme can provide effective security for Intent communication, and it has little effect on the speed and efficiency of the application itself, so that the developers can only focus on their own business logic without worrying about the security problems related to Intent communication.

Key words: Android system, Intent test, decision trees, reinforcement, filter