Techniques and Software for Development and Evaluation of Trading Strategies

This thesis deals with the analysis and development of techniques and software for trading strategies. Within a developed simulation environment algorithms, both existing and newly-introduced, are put to the test and conclusions are drawn regarding both the prediction accuracy and the test procedure as such. The problems with optimization of trading rules are attacked with a non-linear constraint to ensure a minimum number of trading signals for a given time period. The thesis also introduces a novel rank measure that takes into account a large number of securities and grades them according to the relative returns. Linear modeling of the ranks yields competitive prediction results and suggests further research along the same track. In another part of the thesis an algorithm is developed for estimating covarianc