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

Table 2

The result of DL models for lag of 3 minutes and forecasting 5 minutes in advance.

DL models	No. of layers	No. of neurons	Dropout	RMSE
RNN	1	120	0.3	2.459
	2	[60,120]	[0.5,0.1]	2.387
	3	[60,80,40]	[0.4,0.2,0.4]	2.370
LSTM	1	50	0.5	2.355
	2	[100,140]	[0.4,0.2]	2.365
	3	[180,120,20]	[0.3,0.3,0.3]	2.348
GRU	1	150	0.2	2.38
	2	[110,30]	[0.3,0.3]	2.377
	3	[60,150,70]	[0.4,0.3,0.3]	2.398
BI-LSTM	1	90	0.2	2.350
	2	[60,140]	[0.4,0.3]	2.377
	3	[100,120,100]	[0.3,0.2,0.4]	2.415