portfolio optimization in gonum framework

[Pär Lorentz]

Hi,

I have a financial background and am looking into ways to incorporate a statistical go tool for portfolio optimisation purposes. I am reading the book 'Machine Learning with Go' by Daniel Whitenack and he has applied much of the ML techniques by utilising gonum among other go-packages. 

I would like to see if anyone in this community has a financial background and could point me in the direction where and a bit how to use the gonum package in order to solve a non-linear mean-variance optimised portfolio in the classical Markowitz framework. All thoughts and discussions are welcome!

Best,

Pär

[Brendan Tracey]

I don't have a financial background. However, at the moment, you may find http://godoc.org/gonum.org/v1/gonum/optimize/ useful. The standard mean/variance optimization is a convex problem, and while we don't have explicit convex solvers tuned for the problem, our current implementations should work well (as long as your problem isn't huge).

[Pär Lorentz]

Thanks for your reply Brendan! 

I will start digging in-depth in the optimize package of gonum. Are there any plans on branching the optimize package too "purify" the tools towards specific real-life application areas such as finance, physics, etc? I guess the idea is to keep it generic by design so branching may be out of scope, which would make sense, but just curious.