CNN-LSTM-Based Late Sensor Fusion for Human Activity Recognition in Big Data Networks

Table 5

The overall performance of all models for PAMAP2 and RealDisp datasets.

	PAMAP2			RealDisp
Model			Acc			Acc
CNN-EF	83.98	84.48	84.78	89.78	89.87	90.61
CNN-LF	86.85	88.38	88.76	91.21	91.84	91.80
LSTM	87.98	88.29	88.20	88.25	88.89	88.93
CNN-LSTM-EF	88.56	88.95	89.23	91.24	91.95	92.21
CNN-LSTM-LF	92.32	92.15	92.49	92.43	93.78	93.85