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

Figure 6

Confusion matrix of classification results in February 2012 based on ResNet18 using extremely limited labelled data.