Jordan norm in julia

[swsw]

Is there a way to do jordan decomposition or find out the jordan norm of a n*n matrix in julia?

I see the following one but the function jordan_form is not available in SymPy.jl.

Jordan form of a matrix

Is there an equivalent of Matlab’s jordan function, to compute the Jordan canonical form of a matrix? I didn’t find it by tab-completion and in the julia observer didn’t find anything relevant. A version over exact rings would be fine.

Any insight?

[thofma]

Do you have an example of a matrix? If your input is rational, the Jordan form might have proper complex entries. What kind of result do you want? Symbolic (exact) or an approximation?

[jd-foster]

As mentioned in that thread, the numerical analysis properties are problematic in computations: see Jordan normal form - Wikipedia

You might be able to leverage Mathematica through MathLink.jl ?

[swsw]

Thank you! I want appropriamate and the input is rational.

[swsw]

Thank you! I see. That might be a potential direction. So looks julia doesn’t have a direct way to do it.

[Oscar_Smith]

Are you sure you need the Jordan form and can’t use another factorization instead (e.g. eigenvalue/SVD etc)? Jordan normal form is often taught in math classes, but rarely what you want computationally.

[swsw]

Thank you! I see that could a reason why it has not been applied in Julia.