VOLUME 15 NUMBER 2 (July to December 2022)

Vladimir A. Malabanan, Joseph Ryan G. Lansangan, and Erniel B. Barrios*

School of Statistics, University of the Philippines Diliman,
     Quezon City, Philippines

Time series data compiled from different sources often yield varying frequencies, some are measured at higher frequencies, others, at lower frequencies. With data measured over spatial units and at varying frequencies, we postulated a semiparametric spatiotemporal model. This optimizes the utilization of information from variables measured at a higher frequency by estimating its nonparametric effect on the response through the backfitting algorithm in an additive modeling framework. Simulation studies support the optimality of the model over a generalized additive model with aggregation of high-frequency predictors to match the dependent variable measured at a lower frequency. Using quarterly corn production as the dependent variable, the model is fitted with predictors coming from remotely-sensed data (vegetation and precipitation indices), and predictive ability is better compared to the generalized additive models. The model is useful in crop forecasting with inputs from big data sources, an innovative complement to crop production surveys in the generation of official statistics in agriculture.