Examinando por Materia "Time series models"

Mostrando 1 - 2 de 2
  • No hay miniatura disponible
    Ítem
    Anuncios macroeconómicos y mercados Accionarios: El caso Latinoamericano
    (Universidad EAFIT, 2011-12-15) Agudelo, Diego A.; Gutierrez, Angelo
    Do stock markets reflect changes on the macroeconomic fundamentals? . The semi-strong form of the Efficient market hypothesis (HEM - Fama 1970) asserts that stock prices should react immediately to the surprise content on announcements of macroeconomic variables, without predictable over or under reaction. We test this in the six main Latin-American equity markets: Argentina, Brazil, Chile, Colombia, México and Perú, for the announcements of Consumer Price Inflation, Central Bank interest rate, GDP growth, Trade Balance and Unemployment rate. Following Flannery and Protopapadakis (2002), we estimate the effect of the surprises of such announcements, using time series models of conditional volatility, controlling of the exchange rate and international stock markets. We found that the effects on the market returns are significant and with the expected sign only for the CPI in Mexico, for the interest rate in Chile and Colombia, and for Unemployment on those three markets. Moreover, in some cases the stock markets incorporate the announcement with a lag, whereas in others, they react to the announcement rather than to the surprise, in conflict with the HEM. We conclude that the Latin-American stock markets react only partially to the macroeconomic announcements and not fully incorporating the new information in an efficient manner.
  • No hay miniatura disponible
    Ítem
    Modelos de predicción estocástica para bitcoin : una evaluación de métodos y desempeño
    (Universidad EAFIT, 2023) Forero Criollo, Juan Sebastián; Hernández Hernández, Caroline; Cadavil Gil, Alejandro
    This research focuses on forecasting Bitcoin (BTC) prices using statistical models, including LSTM, GRU, SVR, decision trees, Random Forest, and XGBoost. We evaluate their performance in terms of R2, RSME, MAPE, Lin Concordance Coefficient (CCC), and Explained Variance Score—metrics selected for their ability to assess regression models. We utilized BTC closing price data from 2014 to 2023, subjected to preprocessing involving cleaning, optimization, and data engineering. The models, initially unoptimized, were enhanced through hyperparameter tuning and specialized statistical techniques such as cross-validation, L1-L2 regularization, Bayesian and genetic optimization. The results highlight XGBoost as the optimal model with the incorporation of iterative hyperparameter tuning, Bayesian optimization, and nested cross-validation. It achieved outstanding values in all evaluated metrics: RSME of USD 30.45, MAPE of 0.09%, R-squared of 1.0, Lin Concordance Coefficient, and Explained Variance Score of 1.0 in each case.