Abstract

Android malware is software designed specifically to corrupt, attack, damage, harm, disrupt, sneak, or
gain illegal access to anything of value to the device. Android malware growth has been expanding so
dangerously as a result of the advancement and connivance of developing techniques. There are several
ways for security attacks that affect benign apps due to some techniques that include open source (OS)
for developing Android applications and the permission process. To overcome this, applications from the
entire categories are generally examined using static analysis approach to identify benign and malicious
applications. This paper presents the classification of Android benign and malicious apps based on their
application category, and research proposed an application detection comparison of malware and benign
apps using the requested permissions and Application Program Interface (API) this will help to achieve
the best performance of classification models in identifying malicious apps in the same category in
Android applications. By applying feature extraction and selection from the used datasets, the Naïve
Bayes classifier with 10 random tests for the classes of both "Entertainment" and "Personalisation"
attained a high-level of true positive and low-level of false-positive.

Key words: Android, Permissions, APIs, Static, and Machine Learning