Computer aided investment optimization on the world's market

Tato pr áce se se zab ývá problematikou volby optim ální ho portfolia na sv ěto- vém kapitálov em trhu pomoc í Markowitzova modelu s cí lem maximalizovat investor ův zisk. Problematika nestacionarity časov ých řad kurz ů akci í na kapit álov ém trhu je řešena pomocí adaptivn ích metod odhad u. Pr áce d ále nasti ňuje i dal ší mo žnou cestu, jak řešit problematiku nestability odhadu kovarian čn í matice, zalo ženou na teorii n áhodn ých matic. Jedn ím z c íl u t éto pr áce je srovn ání dosa žen ých v ýsledk ů tradi ční m mo- delem a modelem s adaptivn í schopností . K tomuto uúčelu bylo vzhledem k v ýpo četn í n áro čnosti a objemnosti dat vyvinuto a optimalizov áno řešení v programovac ím jazyce MATLAB, kter ý je z ároveň vybaven pot řebn ým matematick ým apar átem pro tuto úlohu.ObhájenoThis diploma thesis deals with the problematics of optimal portfolio choice with focus on the world's stock market. The goal of such an optimization is a maximization of investor's revenue and is based on the Markowitz optimal portfolio model. The problematics of non-stationary time serieses describing stocks on the world's stock market is solved by adaptive parameter estimations. This work introduces another possible way to deal with the phenomenon of non-stable covariance matrix estimates based on random matrix theory. One of the goals of this thesis is comparison of the obtained results - re- sults from traditional model and from its adaptive extension. For this purpose was designed and optimized a software tool using MATLAB as a programming language since the computational intensity and data volume is large and this language is also optimized for mathematical computations