Identifying Malicious Software on Android Devices Through Genetic Algorithm-Driven Feature Selection and Machine Learning

In today’s technology-driven world, the escalating market share of the Android operating system has brought to the forefront an urgent concern surrounding its security flaws. As the platform gains dominance on a global scale, the need for efficiently detecting malware on Android devices has become more critical than ever. One of the main challenges lies in the complex interplay between permissions and application programming interface (API) calls within Android apps. These elements can provide valuable insights into the behavioural patterns of an Android app, but most research studies have narrowly focused on individual permissions or API features. Unfortunately, this approach fails to account for the intricate correlations and patterns hidden within these elements, limiting its effectiveness. A few researchers have attempted to pinpoint combination modes within authorization features that indicate malware presence. However, the results have been far from conclusive, making it challenging to accurately detect malicious activity based solely on these combinations. In light of these obstacles, this article introduces a novel method for effectively identifying Android malware. By combining the strengths of frequent pattern mining and Naive Bayes, this approach offers a promising solution to the malware detection problem. By leveraging the power of these two techniques, we hope to enhance Android security and protect users from potential threats.