View on GitHub

Bearded-android-docs

EigenValues

Download this project as a .zip file Download this project as a tar.gz file

Created Friday 15 November 2013

Let `vec v` be a vector, `alpha` a scalar, and M be a matrix.

`alpha` is an eigenvalue of M and `vec v` is an eigenvector of M if `M vec v = alpha vec v`.

Calculating the Eigenvalues and Eigenvectors

If M is an n x n matrix then `det(M - alpha I)` is an nth-degree polynomial (in `alpha`) called the characteristic polynomial.

To find the eigen values of a matrix, all you need to do is find the roots of the characteristic polynomial.

Example

`M = ((2,7), (-1,-6))`

Then
`M - alpha I`
`= M - ((alpha,0), (0, alpha))`
`= ((2-alpha,7), (-1,-6-alpha))`

So
`det(M - alpha I)`
`= (2-alpha)(-6-alpha) - (-7)`
`= -12+4alpha+alpha^2 + 7`
`= alpha^2 + 4 alpha - 5`
`= (alpha + 5)(alpha - 1)`

The eigenvalues are -5 and 1.


Backlinks:

MachineLearning:NeuralNetworks:Appendix
comments powered by Disqus