Grouped multivariate functional time series method: An application to mortality forecasting

Abstract

Age-specific mortality rates are often disaggregated by different attributes, such as sex and state. Forecasting age-specific mortality rates at the sub-national levels may not add up to the forecasts at the national level. Further, the independent forecasts may not utilize correlation among sub-populations to improve forecast accuracy. Using Japanese mortality data, we extend the grouped univariate functional time series methods to grouped multivariate functional time series forecasting methods.