Research

Machine learning

On my master’s I’m currently studying multidimensional regression with support vector machines. The problem consists in mapping a function from a input vector to a output vector. I’ve also worked with neural networks in time series and bayesian networks in time series.

Macroeconomic simulations

For my Bachalor’s Thesis I’ve studied the Dornbusch Model. The paper was published in 1976 and it’s the most important paper of the author, as long as a important subject in the study of macrointernational economics. It shows, using differential equations and matrix algebra, how the exchange rate depreceates more in the short run than it should accordantly to the exchange rate of equilibrium of the long run. The event is a consequence of the use of the hypothesys of the perfect foresight and the sticky prices. There’s no clear evidence of that effect in real life mainly because there are many factors that changes the exchange rate, but the effect, known as “overshooting”, can be used to help policy makers to understand the effects of a monetary expansion. My work in the model is to implement it in Matlab and, using real data from Brazil, simulate an monetary expansion. The effects will be described and analysed.