Data Anomaly Detection through Semisupervised Learning Aided by Customised Data Augmentation Techniques

Table 2

Dataset composition for the SSL training process.

Labelled/unlabelled data composition of each pattern
Data pattern	Normal	Missing	Minor	Outlier	Square	Trend	Drift
Labelled data	200	200	200	200	200	200	200
Unlabelled data	13,375	2,742	1,575	327	2,796	5,578	479
Total quantity	13,575	2,942	1,775	527	2,996	5,778	679
Total	28,272/100%