Short-Term Reactive Power Forecasting Based on Real Power Demand Using Holt-Winters’ Model Ensemble by Global Flower Pollination Algorithm for Microgrid

Table 16

Comparative analysis of error metrics between the benchmarking model and proposed model.

Performance	Metric	ARIMA	SARIMA	XGBoost	HW-GFPA
Validation	nMBE	0.68	0.61	0.49	0.42
	nMAE	0.830	0.828	0.824	0.822
	nRMSE	0.803	0.804	0.804	0.803
	RMSE	0.700	0.628	0.504	0.432
Testing	nMBE	-0.26	0.48	0.67	0.43
	nMAE	0.802	0.824	0.830	0.840
	nRMSE	0.813	0.805	0.803	0.799
	RMSE	-0.267	0.494	0.690	1.060