Rumor Situation Discrimination Based on Empirical Mode Decomposition Correlation Dimension

Table 6

Comparison of experimental results.

Research work	Feature	Classification method	Evaluation index
			Accuracy (%)
Castill et al. [1]	Content, user, propagation	Decision tree	86	84.9%
Qazvinian et al. [2]	Content, user	Support vector machine (SVM)		89.7%
Yang et al. [3]	Content, user, propagation	Decision tree	78.7
Sun et al. [4]	Content, user	SVM		70.4%
Zhang et al. [5]	Content, user	SVM		74.4%
Liang et al. [6]	User	SVM		76.9%
	User	Decision tree		85.9%
	User	Naive Bayesian model		77.8%
Kwon et al. [8]	Propagation	SVM	87.3	86.7%
	Propagation	Decision tree	82.1	82.2%
	Propagation	Naive Bayesian model	89.7	87.8%
Yuan et al. [33]	Propagation	SVM	85	84%
Wu et al. [17]	Content, user, propagation	Ensemble learning	80
Liu et al. [20]	User	Information dissemination model		81.3%
Ma et al. [21]	Content	Recurrent neural network (RNN)	88.1	89.8%
Liu et al. [23]	User, propagation	Long short-term memory (LSTM)	94.8	94.6%
Srinivasan et al. [26]	Content	Convolutional neural network (CNN)	88	92%
This study	The fractal characteristics of propagation	K-nearest neighbor (KNN)	87.5	87.1%