136
Chapter 9. Signed graph-based semi-supervised learning
(a) Synthetic — two classes (250 and 1000 nodes).
(b) USPS — two classes (250 and 1000 nodes).
(c) ISOLET — two classes (75 300 nodes).
Figure 9.10: Error rates for two classes with the same number of labelled nodes, but
different numbers of nodes in each class
Chapter 9. Signed graph-based semi-supervised learning
(a) Synthetic — two classes (250 and 1000 nodes).
(b) USPS — two classes (250 and 1000 nodes).
(c) ISOLET — two classes (75 300 nodes).
Figure 9.10: Error rates for two classes with the same number of labelled nodes, but
different numbers of nodes in each class