Deep Learning Methods for Malware and Intrusion Detection: A Systematic Literature Review

Table 4

Summary of the metadata extracted from the literature on IoT-based malware detection.

Ref.	Description: method and features used to train and evaluate model	DL algorithm used	Library/framework used	Targeted platform	Dataset; used	Accuracy/F1 score
[100]	IoT and IoBT malware detection through op-code analysis	CNN	Not stated	IoT, IoBT	Self-generated	99.68%
[101]	Behavior-based deep learning framework for detecting malware in IoT environment	SAE	Keras	IoT	Self-generated	About 98.6%
[102]	Detecting pirated software and security threats in internet of things environment	CNN	TensorFlow, Keras	IoT	Not stated	96% for piracy detection, 97.46% for malware detection
[103]	Malware detection in the Internet of things environment by decompiling and extracting op-codes from application samples	LSTM	TensorFlow, Keras, scikit-learn	IoT	Self-generated	98.18%
[104]	Attack detection in industrial Internet of things environment	DAE, DFFNN	Not stated	IoT	NSL-KDD, UNSW-NB15	Up to 98.6%
[105]	Classification of malicious applications in the Internet of things environment by using graph embedding	CNN	Not stated	IoT	Self-generated	88.5%
[106]	Malware detection in industrial IoT devices by extracting DEX file and converting to java class file and extracting the bytecode from the class file	CNN	TensorFlow	IoT	Leopard mobile dataset	98.7%