PCA procedure includes SVD of Covariance matrix. Based on eigenvalues we can find a proportion of variance explained by related Principal Components (eigenvectors).

variance explained by PC1 (in case symmetric matrix) = eigenvalue1 / (sum of all eigenvalues) * 100%

We form a covariance matrix as [ [SD1^2, RO*SD1*SD2], [RO*SD1*SD2, SD2^2]] (SD = standard deviation, RO = Pearson correlation coefficient)

It is clear that if RO = 1 then PC1 explains ~100% of the variance.

Can you please suggest intuition and a general (mathematical) connection between Pearson correlation coefficient and proportion of variance explained in PCA?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.