A Low Computational Cost Method for Mobile Malware Detection Using Transfer Learning and Familial Classification Using Topic Modelling

Table 8

Comparative analysis of various Android malware detectors on the basis of objective, clustering, robustness, and computational cost.

Technique/Method	Year	Approach	Objective	Familial clustering	Detection robustness				Computational cost
					Reflection	Resource obfuscation	System call obfuscation	Dynamic permissions
DroidCat [25]	2018	Dynamic	Detection and categorization	✓	✓	✓	✓	✓	N/A
AOMDroid [41]	2020	Dynamic	Detection	✕	✕	✓	✓	✓	High
MamaDroid [33]	2017	Static	Detection	N/A	✕	✕	✓	✓	N/A
DroidSieve [26]	2017	Static	Detection and categorization	✓	✓	✕	✓	✕	N/A
Android-SEM [48]	2022	Dynamic	Detection and categorization	✓	✕	✕	✓	✕	Moderate
DroidScribe [49]	2016	Dynamic	Categorization	✓	✓	✕	✕	✕	Moderate
MalDozer [11]	2018	Dynamic	Detection and categorization	✓	✕	✓	✓	✓	High
DL-Droid [50]	2020	Static and dynamic	Detection	✕	✕	✕	✕	✓	High
Ad-Mat [39]	2021	Static	Detection and categorization	✓	✕	✕	✓	✕	High
Proposed method	This study	Static and dynamic	Detection and categorization	✓	✓	✕	✓	✓	Low