Structural Vibration Data Anomaly Detection Based on Multiple Feature Information Using CNN-LSTM Model

Table 4

The calculation results of the different models.

Model	Accuracy (%)	Loss	Time (s)
			Preprocessing	Training	Total
1D-CNN	88.78	0.5242	238.8	21.4	260.2
CNN	89.43	0.3740	583.3	54.9	638.2
LSTM	89.02	0.4158	238.8	86.6	325.4
Bi-LSTM	94.46	0.1256	238.8	200.0	438.8
CNN-LSTM	99.83	0.0140	238.8	37.8	276.6