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

Algorithm 1

Dominant feature selection.

	Input: Unique feature vector data for both clean and ransomware samples
	Output: List of k-dominant features
	Symbols Used: Let Sc be the total number of clean sample, Sm be the total number of ransomware samples, and K be the number of dominant features required to be extracted
	Step 1: for all clean samples, calculate sum of frequencies of each feature and normalize it
	Step 2: for all ransomware samples, calculate sum of frequencies of each feature and normalize it
	Step 3: for all features in unique feature list, calculate the absolute difference between normalized frequencies of clean and ransomware sample
	Step 4: Sort
	Step 5: Choose k to record k number of dominant features for k in (20, 40, 60, 80) iteratively.