Long-Term Yield Prediction of Greenhouse Sweet Pepper Crops

Abstract

In this paper, a new model for predicting the yield of greenhouse sweet peppers (Capsicum annuum L.) is presented. The model can provide long-term prediction up to 7 weeks in advance with the same accuracy it can predict yield one week in advance. It uses both past and expected environmental readings as well as physiological data as input to a specially designed artificial neural network. The model was tested using 4 years of data that was obtained from commercial pepper growers. Short-term prediction accuracy (one week) is consistent with other predictive models in the literature for sweet peppers. This validates our long-term results.