Real-Time System Prediction for Heart Rate Using Deep Learning and Stream Processing Platforms

Table 3

The result of DL models for lag of 5 minutes and forecasting 7 minutes in advance.

DL models	No. of layers	No. of neurons	Dropout	RMSE
RNN	1	100	0.4	2.887
	2	[200,40]	[0.5,0.1]	2.878
	3	[100,140,60	[0.5,0.4,0.4]	2.876
LSTM	1	100	0.4	2.822
	2	[50,130]	[0.5,0.1]	2.825
	3	[170,50,70]	[0.4,0.4,0.4]	2.862
GRU	1	50	0.3	2.833
	2	[200,140]	[0.3,0.3]	2.831
	3	[140,50,180]	[0.3,0.3,0.3]	2.817
BI-LSTM	1	80	0.4	2.823
	2	[220,80]	[0.5,0.4]	2.931
	3	[120,100,180]	[0.2,0.3,0.3]	2.810