Online Semisupervised Learning Approach for Quality Monitoring of Complex Manufacturing Process

Table 1

List of notations.

Notation	Meaning
	data batch in data streams
	Single input data vector and single ground truth output vector separately
	Input data batch and batch label
	The current network parameters and optimal network parameters
	Network parameter importance, calculated by (3)
	Regularization factor:
,	Predefined thresholds in (2)
	The cardinality of the cluster
	The cardinality of the class of the cluster
	Convolution layer
	Feature map
	Cluster center
	Partially destroyed input vector with the masking noise
	The distance between two data samples
	The contribution of cluster
	The mixing coefficient for hidden node pruning criterion
	Reconstructed symbol
ACM	Autonomous clustering mechanism
ori	Original data
aug	Augmented data (generation of augmented label of Section 3)
ps	Pseudodata (generation of pseudolabel of Section 3)