Edge Caching in Fog-Based Sensor Networks through Deep Learning-Associated Quantum Computing Framework

Table 5

Comparison with previous research.

Research	Deployed algorithm	Performance measurements	Results
[39]	Multiagent deep reinforcement learning (MADRL)	Caching reward	21%
		Cache hit rate	Highest
		Traffic load	43%
	Multiagent actor-critic (MAAC)	Caching reward	56%
		Cache hit rate	Higher
		Traffic load	45%
	Deep reinforcement learning (DRL)	Caching reward	43%
		Cache hit rate	Lower
		Traffic load	36%
	Least recently used (LRU)	Caching reward	34%
		Cache hit rate	Lowest
		Traffic load	12%
[40]	Personalized edge caching system (PECS)	Deep packet inspection	Top-down analysis (network level) and bottom-up analysis (user level)
[41]	One-dimensional convolutional neural network (ODCNN) Self-Organizing Map (SOM)	Accuracy rate	99.8%
[42]	Support vector machine	Accuracy	0.984
		Precision	0.984
		Recall	0.983
		F1 score	0.981
	Logistic regression	Accuracy	0.983
		Precision	0.982
		Recall	0.983
		F1 score	0.983
	K-nearest neighbors	Accuracy	0.984
		Precision	0.983
		Recall	0.984
		F1 score	0.984
	Isolation forest	Accuracy	0.870
		Precision	0.969
		Recall	0.973
		F1 score	0.919
Proposed	Self oranizing map (SOM)	Quantization error	0.000024
		Topographic error	0.092
		QE + TE	0.0000235
	Two-level spin quantum phenomenon (TLSQP)	Basic states	0	OV = 85.4%	PV = 85.2%
			1	OV = 14.6%	PV = 14.8%