Functional Time Series Prediction Under Partial Observation of the Future Curve

–This article tackles one of the most fundamental goals in functional time series analysis which is to provide reliable predictions for future functions. Existing methods predicting a complete observation use only completely observed trajectories. We develop new method, called partial prediction (PFP), uses both trajectories and information (available data) on trajectory be predicted. The PFP method includes an automatic selection criterion tuning parameters based minimizing error, convergence rate established. Simulation studies demonstrate that incorporating partially outperforms existing with respect mean squared error. illustrated superior environmental data traffic flow data.