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

Table 6

Confusion matrix for the PAMAP2 Dataset using CNN-LSTM-LF.

	Lying	Sitting	Standing	Walking	Running	Cycling	Nordic walking	Ascending stairs	Descending stairs	Vacuum cleaning	Ironing	Rope jumping
Lying	97.6	0.37	0.67	0.05	0.04	0.02	0.03	0.27	0.15	0.52	0.29	0.01
Sitting	0.36	95.1	1.47	0.01	0	0.15	0.15	0.23	0.16	0.88	1.38	0.07
Standing	0.17	0.25	95.9	0.15	0.03	0.02	0.26	0.25	0.36	0.64	1.94	0.02
Walking	0.05	0.06	0.73	94.3	0.02	0.04	2.01	1.04	1.02	0.34	0.34	0.06
Running	0.11	0.06	1.03	0.6	95.4	0.05	0.39	0.51	0.45	0.36	0.49	0.57
Cycling	0.02	0.09	0.06	0.11	0	95.5	0.25	0.24	0.45	1.58	1.64	0.04
Nordic walking	0.02	0.04	0.10	2.81	0.04	0.06	94.6	0.60	0.36	0.42	0.8	0.08
Ascending stairs	0.77	0.33	2.24	4.13	0.17	0.48	2.07	81.2	4.68	2.4	1.15	0.38
Descending stairs	0.36	0.29	2.00	4.18	0.19	0.74	1.20	3.33	82.6	2.31	2.01	0.81
Vacuum cleaning	0.22	0.27	0.88	0.34	0.02	2.05	0.64	1.17	0.95	84.7	8.69	0.06
Ironing	0.10	0.29	2.05	0.06	0.03	0.32	0.18	0.13	0.26	2.47	94.1	0.06
Rope jumping	0.22	0.39	0.82	0.82	0.48	0.45	1.51	0.84	2.94	2.76	3.38	85.4