Table 1.

Different interpretations of the eigenvectors and eigenvalues resulting from CA.

Method	Interpretation eigenvectors	Interpretation eigenvalues
Gradient analysis using canonical correlation analysis	Latent variable	Strength of correlation between row and column scores
Graph partitioning using the normalized cut	Approximate cluster labels	Quality of the partitioning (given by normalized cut)
Dimensionality reduction using graph embedding	Coordinates in the embedding space	Variation explained