A Hybrid LSTM-Based Ensemble Learning Approach for China Coastal Bulk Coal Freight Index Prediction

Table 8

Improvement percentage of predictive performance with adding demand and demand-related factors.

Evaluation criteria	Models	Improvement percentage of monthly forecasting (%)
MAPE	LSTM	12.67
	GBRT	9.97
	LSTM-GBRT	5.32
	RF	9.81
	LSTM-RF	5.16
RMSE	LSTM	10.01
	GBRT	8.90
	LSTM-GBRT	5.18
	RF	7.87
	LSTM-RF	5.01