Dominant Feature Selection and Machine Learning-Based Hybrid Approach to Analyze Android Ransomware

Table 1

Comparison with existing studies.

Reference	Approach	Machine learning model used	Feature set used	Accuracy (%)
[2]	Static	Random forest, logistic regression, XGBoost, Naive Bayes, support vector machine (SVM), deep learning, and decision tree classifier	Intent, permission, API calls, system commands, and malicious activities	96.3
[3]	Dynamic	Naive Bayes, SVM, and logistic regression	Application programming interface (API)	97
[4]	Hybrid	SVM	Permission, API calls, system calls	99.7
	Our proposed method (dynamic)	J48, LMT, random forest, and random tree	System calls, system components, system command, phone events, run-time permissions, and broadcast receivers	99.85