What is the best way to implement time series statistical methods, for example ARMA models, ARCH, kernel regressions or even a neural network? I have programming skills in C++, but I have not ventured past standard AFL. I'm guessing (thought not entirely sure) that I could used a plug-in, run a script or just abandon amibroker for this and use another package.
I would prefer to have all of my analysis, routines, etc., to be streamline and run as efficiently as possible, e.g. without having to copy and paste information. Has anybody else done this sort of thing before? What would you suggest? Thanks, JH
