A0246
Title: Nonlinear prediction of functional time series
Authors: Haixu Wang - University of Calgary (Canada) [presenting]
Abstract: A nonlinear prediction (NOP) method is proposed for functional time series. Conventional methods for functional time series are mainly based on functional principal component analysis or functional regression models. These approaches rely on the stationary or linear assumption of the functional time series. However, real data sets are often non-stationary, and the temporal dependence between trajectories cannot be captured by linear models. Conventional methods also make it hard to analyze multivariate functional time series. To tackle these challenges, the NOP method employs a nonlinear mapping for functional data that can be directly applied to multivariate functions without any preprocessing step. The NOP method constructs feature space with forecast information; hence, it provides a better ground for predicting future trajectories. The NOP method avoids calculating covariance functions and enables online estimation and prediction. The finite sample performance of the NOP method is examined using simulation studies that consider linear, nonlinear, and non-stationary functional time series. The NOP method shows superior prediction performances in comparison with the conventional methods. Three real applications demonstrate the advantages of the NOP method model in predicting air quality, electricity price and mortality rate.
