Remaining Useful Life Estimation through Deep Learning Partial Differential Equation Models: A Framework for Degradation Dynamics Interpretation Using Latent Variables

Table 3

Machine learning classifiers performance with TH = 50 for the FD004 subdataset.

ML classifier	Set	Accuracy	False negative	False positive	F1-score (%)	Recall (%)	Training time (s)
Nearest neighbors	Train	92.4	4.9	2.7	81.9	77.8	4.02
	Validation	87.8	7.2	5.0	71.2	67.7
Linear SVM	Train	89.0	8.3	2.7	71.3	62.3	94.96
	Validation	89.1	8.2	2.6	72.1	63.0
Random forest	Train	89.1	7.9	3.0	72.1	64.0	0.98
	Validation	89.2	7.9	2.8	72.8	64.5
Neural network	Train	89.0	7.5	3.5	72.6	66.0	9.31
	Validation	89.2	7.5	3.3	73.3	66.5
AdaBoost	Train	88.8	7.8	3.5	71.7	64.7	4.82
	Validation	88.9	7.8	3.3	72.4	65.1
Logistic regression	Train	89.0	8.0	2.9	71.7	63.4	1.53
	Validation	89.2	8.0	2.8	72.6	64.1