Applying Bayesian probability for Android malware detection using permission features

• Main Authors: Sharfah Ratibah, Tuan Mat, Mohd Faizal, Ab Razak, Mohd Nizam, Mohmad Kahar, Juliza, Mohamad Arif, Azlee, Zabidi
• Format: Conference or Workshop Item
• Language: English
• Published: IEEE 2021
• Subjects: QA75 Electronic computers. Computer science
• Online Access: http://umpir.ump.edu.my/id/eprint/32537/1/Applying%20Bayesian%20probability%20for%20Android%20malware.pdf; http://umpir.ump.edu.my/id/eprint/32537/; https://doi.org/10.1109/ICSECS52883.2021.00111
• Institution: Universiti Malaysia Pahang

he tremendous rise of mobile technology has boosted malware and has raised the threat of malware. The proliferation of malware has given a great concern among mobile users. Various approaches have been applied to prevent malware spread, including firewalls, antivirus software and many more methods. Google has provided permission features as the main security to filter out the possibility of malware-infected Android mobile. Nevertheless, some permissions immediately granted by Android without user confirmation. This paper proposes a malware detection system based on permission features using Bayesian probability to battle the malware issue. This study used 96,074 samples retrieved from Androzoo and Drebin. By using static analysis, this study focuses on permission features that are significant in Android applications. The experiments conducted using chi-square as an algorithm and Naïve Bayes as a classifier. The accuracy of the detection is 85%. In conclusion, the detection of Android malware using the dataset has produced a good performance.