Buscador Google

Maestría en Finanzas Cuantitativas


Arlington Fonseca Lemus - Graduate​ (08/2018)
Title: Wavelet analysis on financial time series

Wavelet methods possess some features that make them a tool with great potential for financial research. The purpose of this thesis is to study the usefulness of wavelet methods in financial time series analysis, for which data from Colombian financial market has been used.
In this thesis the wavelet theory is briefly presented, with a special focus on the Discrete Wavelet Transform and Daubechies wavelets. Then, a multiresolution decomposition is illustrated for two distinct log-returns series. Finally, a wavelet-based prediction approach is presented, as well as a comparison between its results and those of a traditional prediction method.
Keywords: Wavelet analysis, Discrete Wavelet Transform, financial time series, multiresolution decomposition, prediction.

Darwin Javier Fonseca Lemus - Graduate (08/2018)
Title: Stock market trading trhough a Neural Network adaptation

In this thesis, several machine learning techniques are explored and applied to the investment decision problem. Some indicators of technical analysis are used as inputs of a neural network which is trained to determine, with each input vector, a buy, sell or hold signal. Likewise, the structure of the network is optimized by applying a genetic algorithm, in order to determine its adequate depth. With this work, some bases are established to perform different empirical studies that improve and deepen the analyzed topics.
Keywords: Artificial intelligence, machine learning, artificial neural network, genetic algorithm, perceptron, technical analysis, decision rules.

Sebastian Mendoza - Graduate (08/2018)
Title: Black Litterman applied to global fixed income securitie

The purpose of this research is to develop the Black Litteman asset allocation model in an intuitive way, showing its most important properties with data from global fixed income securities. The model begins with a market portfolio, which the investor adjusts with his views on securities return. Results show that adjusting the market portfolio without the Black Litterman model produces incoherent portfolios. When views are incorporated with the Black Litterman model, the optimal portfolio reflects these initial intentions. The investor can also adjust his confidence in his views, which will also be captured in the resulting optimal portfolio. Our demonstration finishes with an analysis of the turnover of consecutive portfolios obtained by Black Litterman vs Markowitz, we find that portfolios using Black Litterman are more stable in time than those estimated with the Markowitz approach, reducing transactional and operational cost for the investor.