Analysis of Feature Selection Methods for Android Malware Detection Using Machine Learning Techniques
Santosh Kumar Smmarwar, Govind Prasad Gupta, Sanjay Kumar · 2023
The simplicity, user-friendly environment, and popularity of Android phones have increased the risk of cyber-attacks on Android devices. A variety of applications containing malicious code are rapidly emerging. Quickly and accurately identifying these applications and them from damaging the privacy and confidentiality of users is a very difficult task because of the rapid availability of new variants of malware. Most of the approaches of machine learning and antivirus software work on predefined pattern sets and have become incapable of detecting the new malware variants, as the performance of pattern-learning methods is greatly affected by appropriate features in detecting malware. The methods of efficient feature selection play a major role in filtering unwanted features. In this chapter, we have shown an analysis of different commonly used feature selection methods used in Android malware detection and evaluated the assessment of the feature selection methods in regards to the accuracy achieved. The accuracy of each classifier was evaluated to identify the most reliable classifier for malware detection using various feature sets. The results of the tests confirm that feature selection approaches are needed to improve the accuracy of the learning models.