Scientific Reports

Table 9 Comparison of R² values for PM_2.5 prediction across stations using deep learning models with unencoded and encoded features.

From: Deep learning framework for hourly air pollutants forecasting using encoding cyclical features across multiple monitoring sites in Beijing

	Unencoded features				Encoded features
	DNN	ANN	CNN	LSTM	DNN	ANN	CNN	LSTM
Aotizhongxin	0.948	0.869	0.948	0.937	0.948	0.866	0.948	0.940
Changping	0.943	0.860	0.943	0.920	0.943	0.817	0.943	0.933
Dongsi	0.952	0.880	0.953	0.938	0.954	0.806	0.953	0.939
Guanyuan	0.955	0.935	0.956	0.944	0.956	0.841	0.956	0.947
Huairou	0.940	0.889	0.940	0.940	0.939	0.806	0.941	0.933
Nongzhanguan	0.961	0.783	0.961	0.944	0.962	0.828	0.961	0.928
Shunyi	0.952	0.867	0.952	0.936	0.953	0.819	0.951	0.938
Tiantan	0.959	0.878	0.959	0.942	0.960	0.850	0.961	0.944
Wanliu	0.957	0.893	0.957	0.940	0.957	0.826	0.957	0.947
Wanshouxigong	0.948	0.869	0.948	0.937	0.948	0.866	0.948	0.940

Back to article page

Search

Advanced search

Quick links