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

Table 2

Machine learning classifiers performance with TH = 50 for the FD001 subdataset.

ML classifier	Set	Accuracy	False negative	False positive	F1-score	Recall	Training time
Nearest neighbors	Train	93.0	4.6	2.4	86.3	82.7	3.57
	Validation	88.7	6.7	4.6	78.1	75.0
Linear SVM	Train	90.6	6.2	3.2	81.3	76.8	69.37
	Validation	90.7	6.1	3.2	81.6	77.2
Random forest	Train	90.7	6.7	2.6	81.1	74.9	0.81
	Validation	90.6	6.7	2.7	81.0	75.1
Neural network	Train	90.6	6.1	3.3	81.3	77.1	10.43
	Validation	90.7	6.0	3.3	81.7	77.5
AdaBoost	Train	90.6	6.5	3.0	81.0	75.7	3.80
	Validation	90.5	6.5	3.0	81.1	75.9
Logistic regression	Train	90.5	6.0	3.5	81.3	77.6	1.44
	Validation	90.6	5.9	3.5	81.6	77.9