cHybriDroid: A Machine Learning-Based Hybrid Technique for Securing the Edge Computing

Table 1

A summary of related work.

References	Methodology				Used feature
	Static	Dynamic	Hybrid	ML-based	Static	Dynamic	Data set
Feizollah et al. [10]	✗	✗	✗	✗	Permission	✓	Custom | Drebin
Almin et al. [5]	✓	✗	✗	✗	Intents	✓	Custom
Canfora et al. [8]	✗	✓	✗	✓	✗	System calls	Custom | Drebin
Wong et al. [4]	✗	✓	✗	✗	✗	Malware tracking through input Genera-	Custom | Drebin
Youngjoon et al. [11]	✗	✓	✗	✗	✗	API calls	Custom
Alzaylaee et al. [2]	✗	✓	✗	✗	✗	API calls	Malgenome data set
Zhao et al. [12]	✗	✓	✓	✓	Permissions	General dynamic activities trigged by	Custom
Dash et al. [13]	✓	✓	✗	✓	✗	System calls, Decoded binder communication, abstracted behavioural patterns	Custom
Xu et al. [14]	✗	✓	✓	✓	Collect attack tree path	Graph kernels	Custom
Yuan et al. [15]	✓	✓	✓	✓	Permissions, sensitive API	DexClass, receive net service start	Custom
Wang et al. [16]	✓	✗	✗	✓	Permissions, API calls, hardware features, code patterns	✗	Custom
Kim et al. [17]	✓	✗	✗	✓	Opcode, API, permissions, component, and environmental and string features	✗	Custom | malgenome data set
Karbab et al. [18]	✓	✓	✗	✓	API method calls	✗	Drebin | malgenome | virushare | contagio minidump
Arshad S et al. [19]	✓	✓	✓	✓	Hardware components requested per missions, application components, and API calls.	System calls	Drebin
Hou et al. [20]	✗	✓	✗	✓	Linux kernel system calls	✗	Custom
Pektas and Acarman [21]	✓	✓	✗	✓	Permissions and hidden payload	API calls, installed services, network connections	Virushare