Skip to main content
Figure 2 | BMC Bioinformatics

Figure 2

From: Spectral embedding finds meaningful (relevant) structure in image and microarray data

Figure 2

Scatter plots of image ordering for six projection methods. (a) Regression of ranked eigenfunction values calculated by PCA-correlation vs. actual ordering. (b) Regression of ranked eigenfunction values calculated by PCA-covariance vs. actual ordering. Note that the scores have been reverse sorted for consistency with the other plots. (c) Regression of ranked eigenfunction values calculated by nonmetric MDS vs. actual ordering. (d) Regression of ranked eigenfunction values calculated by kernel PCA (ε = 1,040) vs. actual ordering. (e) Regression of ranked eigenfunction values calculated by the spectral method from Ng et al. (ε = 70) vs. actual ordering. (f) Regression of ranked eigenfunction values calculated by the spectral method from Lafon [3,4] vs. actual ordering.

Back to article page