Before part III of forecasting with RSS+SVM+wavelets I thought it would help to give some useful concepts from Hilbert space. Attached is a very rough look, and also an application using orthogonal functions to model a stationary signal (Fourier series). This will contrast nicely with wavelets, which are most useful for non-stationary signals eg., stock indices.
hilbert space overview
Fourier series example
Quick tour through Hilbert space
February 4, 2009
Leave a Comment » | 
quantum mechanics, stochastic differential equations | Tagged: Hilbert, Fourier series | 
Permalink
Posted by bbrouwer
SDE scripts
January 19, 2009Mostly a reworking of fortran code from Kloeden & Platen into Matlab/Octave. Includes important things like Karhunen-Loéve expansions, Stratonovich integrals for higher order methods, Ito summation etc and also a Markov Chain Monte Carlo example
Leave a Comment » | 
Monte Carlo, stochastic differential equations | Tagged: Ito, Karhunen Loeve, Markov Chain Monte Carlo, Stratonovich | 
Permalink
Posted by bbrouwer
