Local Gravitation Clustering-Based Semisupervised Online Sequential Extreme Learning Machine

Figure 1

An example of the empirical distribution mismatch between labeled and unlabeled samples with the Jain’s toy data. The labeled and unlabeled samples are displayed in the (A) of (b) and (c), and the kernel density estimations of their x-axis projection are plotted in the (B) of (b) and (c), respectively. (a) Original samples. (b) Labeled samples S^l. (c) Unlabeled samples (S^u).

(a)

(b)

(c)