Pruning Multilayered ELM Using Cholesky Factorization and Givens Rotation Transformation

Figure 5

Comparison of training time between the algorithms MELM and P-MELM using (a) Diabetic Retinopathy Debrecen, (b) image segmentation, and (c) coal spectral data, while the hidden node is decreased one by one.

(a)

(b)

(c)