A1561
Title: Similarity-based path forecasting of U.S. recession periods
Authors: Henri Nyberg - University of Turku (Finland) [presenting]
Visa Kuntze - University of Turku (Finland)
Samuel Rauhala - University of Turku (Finland)
Abstract: A nonparametric similarity-based approach is developed to obtain path forecasts for binary time series by finding probability forecasts for each viable sequence of observations multiple periods ahead. In contrast to the common way of specifying forecast horizon-specific parametric models, the path forecasts are internally consistent and obtained simultaneously for all the horizons. In an empirical illustration, the state of the U.S. business cycle is forecasted around the past three recession periods.
