A Novel Carbon Price Fluctuation Trend Prediction Method Based on Complex Network and Classification Algorithm

Table 1

Summary of studies on price forecasting via various methods.

Types	Typical literature	Forecasting models	Data type	Main results
Econometric models	Lanza et al. [7]	ECM	Daily	The cointegration marginally improves static forecasts
	Murat et al. [8]	VECM	Weekly	VECM outperforms the random walk model (RWM)
	Baumeister et al. [9]	VAR	Monthly	VAR tends to have lower MSPE at short horizons than the AR and ARMA models
	Xiang Y [11]	ARIMA	Daily	ARIMA model possesses a good prediction effect and can be used as a short-term prediction
	Fan et al. [12]	GED-GARCH	Daily	GED-GARCH model has superior power in the out-of-sample forecast compared with the popular HSAF method
	Mohammadi et al. [13]	GARCH, EGARCH, APARCH, and FIGARCH	Weekly	APARCH model outperforms other models
AI models	Atsalakis et al. [14]	ANN	Daily	ANN model outperforms the ARMA and GARCH models
	Yahşi et al. [15]	GEP	Daily	GEP model outperforms the ANN and ARIMA models
	Xie et al. [16]	SVM	Monthly	SVM model outperforms the ARIMA and BPNN models
	Zhu et al. [17]	LSSVM	Daily	LSSVM forecasting model outperforms the SVM and RBF network models
	Lu et al. [18]	MML	Monthly	MML model has superior power compared with the traditional machine learning models
Hybrid models	Zhu et al. [19]	GMDH-PSO-LSSVM	Daily	GMDH-PSO-LSSVM model performs better than the conventional LSSVM model
	Gao et al. [20]	EMD-PSO-SVM	Daily	EMD-PSO-SVM model is significantly superior to the single SVM model
	Zhu et al. [21]	Hybrid ARIMA and LSSVM methodology	Daily	ARIMA-LSSVM models outperform their single benchmarks in both level and directional predictions
	Huang et al. [27]	PSO-RBF	Monthly	PSO-RBF approach is able to improve prediction accuracy and to simplify the complexity of the RBF model structure
	Liu et al. [29]	EWT-GRU	Daily	The proposed approach is significantly effective and practically feasible