Hey everyone,

So I have three eigenvalues with respective confidence intervals...

e1 = 10.9922 +/- 1.36
e2 = 4.2418 +/- 0.52
e3 = 0.7660 +/- 0.09

e1 + e2 + e3 = N

However, in my research (geology) they are more useful in a normalised form (Si = ei/N) so that S1 + S2 + S3 = 1.

This is where my knowledge of stats falls short! Is there a way of propagating the CI values to a normalised form as well?

Cheers