Today, there is a large amount of assets which are offered to investors, and if we consider the possibility of relocating the investor's funds, we come to a very complicated problem, which this thesis aims to cover. The main objective is to explore the basics of the portfolio theory and its real usage in practice. Emphasis is put on the periodic re-optimizing of the investor's portfolio and getting the answer to the question of whether such conduct is more successful than the standard and investment in equity indices. The theoretical part summarizes the currently used approaches to optimization which are tested in the practical part on real data and evaluated.

Submitted by Verena Pereira (verenagoncalves@uefs.br) on 2018-07-20T23:30:41Z No. of bitstreams: 1 Disserta??oMestradoVers?oFinalCD.pdf: 2552948 bytes, checksum: 021c9ef4954fba969e5860759364257e (MD5) / Made available in DSpace on 2018-07-20T23:30:41Z (GMT). No. of bitstreams: 1 Disserta??oMestradoVers?oFinalCD.pdf: 2552948 bytes, checksum: 021c9ef4954fba969e5860759364257e (MD5) Previous issue date: 2018-03-06 / The biggest challenge for stock market traders is to identify the timing to enter and exit in a trade. This research uses a trading system based on bullish and bearish candlesticks patterns to identify buy and sell signals in the brazilian?s stock market with exit through the chandelier method. The trading system was tested during the period 2005 and 2010 for application in two strategies between 2011 and 2016. The statistical significance and robustness of the strategies were evaluated through skewness, kurtosis, Monte Carlo, z-score, t-test and walk-forward. They revealed some prediction in bullish candlesticks standards / O maior desafio para os investidores no mercado de a??es ? identificar o momento para entrar e sair de uma negocia??o. Esta pesquisa utilizou um sistema de negocia??o baseado em padr?es candlesticks de alta e de baixa para gerar sinal de compra/venda no mercado brasileiro de a??es e vender/comprar por meio do m?todo chandelier exit. O sistema de negocia??o foi testado para simular negocia??es no per?odo entre 2005 e 2010 e para aplica??o em duas estrat?gias no per?odo entre 2011 e 2016. A signific?ncia estat?stica e robustez das estrat?gias foram avaliadas por meio daskewness, kurtosis, Monte Carlo, z-score, t-test e walk-forward. Eles revelaram algum grau de predi??o nos padr?es de candlesticks de alta

This thesis is devoted to the technical analysis with the emphasis on design, testing and using of trading systems. Its objective is to find out whether it is possible for a trader to design and trade his own profitable trading system with widely accessible tools and methods. First part of the thesis is focused on the chart analysis and description of candlestick charts including their rate of profit success, all based on hands-on experience in a real market. It continues with a breakdown of most used methods based on moving averages. The second chapter fully describes main stage of trading system development and follows up with third chapter on practical application of the theoretical assumption on the real market conditions, i.e. to design a profitable trading system

This study is a comparison of forecasting methods for predicting the daily maximum air temperatures in Uppsala using real data from the Swedish Meteorological and Hydrological Institute. The methods for comparison are univariate time series approaches suitable for the data and represent both standard and more recently developed methods. Specifically, three methods are included in the thesis: neural network, ARIMA, and naïve. The dataset is split into a training set and a pseudo out of sample test set. The assessment of which method best forecast the daily temperature in Uppsala is done by comparing the accuracy of the models when doing walk forward validation on the test set. Results show that the neural network is most accurate for the used dataset for both one-step and all multi-step forecasts. Further, the only same-step forecasts from different models that have a statically significant difference are from the neural network and naïve for one- and two-step forecasts, in favor of the neural network.

This diploma thesis deals with the design of a trading strategy and subsequent implementation of an automated trading system for the forex currency market. In this thesis, a "breakout" strategy with trade filtering based on moving average is created. Consequently, an automated trading system for the MetaTrader 4 platform is developed in MQL4 language. This thesis also deals with the back-testing and optimization of the system in order to maximize the stability and profit.