Python eigenvectors

Question

eigenvalues, eigenvectors = linalg.eig(K)

How can I print just eigenvectors of len(K). So if there is K, 2x2 matrix, I get 4 eigenvectors, how can I print just 2 of them if there is len(K)=2....

Many thanks

Swanskin

why not print eigenvectors[:len(K)] ? – Pugliese

This matrix K is not always 2x2....it's random 4x4,.... – Swanskin

Also I can add, print eigenvectors[:len(K)], this will not work if there is just 2x2 matrix, it print's 4 elements, not just 2.... – Swanskin

@thaking: Four vectors? You are probably getting two vectors each of length two (four numbers in total). Can you show the exact matrix you're passing to eig and the exact results you're getting back. – Garmon

Yes 2 vectors length two; 4 numbers TOTAL; I only need on vector, if there is K 2x2; – Swanskin

how can you get 4 eigenvectors from 2x2 matrix??? – Personalism

@Swanskin Slow down cowboy. You don't understand eigenvectors. Head back to your maths text books before proceeding. – Windhoek

Ok, it seems like I didn't get it. I thought you should write print eigenvectors[: eigenvalues.index(<index of first almost zero eigenvalue>)] – Pugliese

I misunderstood something in my code, so i was little confused,sorry, I solved my problem now. – Swanskin

Answer 1

You are getting two vectors of length two, not four vectors. For example:

In [1]: import numpy as np

In [2]: K=np.random.normal(size=(2,2))

In [3]: eigenvalues, eigenvectors = np.linalg.eig(K)

In [4]: eigenvectors
Out[4]: 
array([[ 0.83022467+0.j        ,  0.83022467+0.j        ],
       [ 0.09133956+0.54989461j,  0.09133956-0.54989461j]])

In [5]: eigenvectors.shape
Out[5]: (2, 2)

The first vector is eigenvectors[:,0], the second is eigenvectors[:,1].

Garmon

Yes, but how to print just 1 vector ? – Swanskin

Thanks I misunderstood something in my code, so i was little confused, thanks. – Swanskin

eigenvectors are columns, not rows; – Personalism

Answer 2

From the manual:

The normalized eigenvector corresponding to the eigenvalue w[i] is the column v[:,i].

Personalism