Predictive modelling with panel data and multivariate adaptive regression splines: case of farmers crop delivery for a harvest season ahead
Stochastic Environmental Research and Risk Assessment
Springer Berlin Heidelberg
This paper investigates a harvest-season level unbalanced panel data (PD) of farmers crop delivery for monitoring the gathering activity and for aiding to support reception and storage decisions making of an agricultural cooperative. To achieve these purposes, the fitting and the prediction of the daily farmers crop delivery quantities were realised based-on the total expected quantity of the whole harvest season, the daily volume of precipitation and the amount of sunshine. In order to capture and extrapolate data patterns, both the PD regression and the multivariate adaptive regression approaches were implemented and tested for a real life agricultural cooperative case study. The obtained results exhibit an accurate predictive modelling of the farmers crop delivery behaviour for harvest seasons ahead.