Attention Mechanism-Based CNN-LSTM Model for Wind Turbine Fault Prediction Using SSN Ontology Annotation

Table 7

Experimental results after feature screening.

Algorithm	Wind turbine icing fault dataset				Wind turbine yaw fault dataset
	No. 21 wind turbine				No. 4 wind turbine
	CLA	0.7776	0.7899	0.7565	0.7728	0.8622	0.8228	0.9234	0.8702
LSTM	0.7664	0.7310	0.7431	0.7531	0.8401	0.8025	0.9023	0.8495
RNN	0.6855	0.6571	0.7756	0.7115	0.7751	0.7912	0.7475	0.7687
XGBoost	0.7345	0.7286	0.7571	0.7426	0.7345	0.7286	0.7571	0.7426