Prediction of pepper (Capsicum annuum L.) leaf area using group method of data handling-type neural networks

Leaf area is a common evaluation applied in experiments on fruit physiology and a routine measurement in experiments investigating horticultural crops studying physiological phenomena such as light, photosynthesis, respiration, plant water consumption and transpiration. Evaluations of leaf area can be time consuming and require sophisticated electronic instruments. However Arti cial Neural Networks present a powerful tool for system modeling. One of the sub models of arti cial neural networks is the group method of data handling-type neural networks (GMDH-type NN). The objective of this research was to develop a simple, accurate, non-destructive and time saving predictive model to estimate leaf area (LA) in pepper. Results suggest that GMDH-type NN provides an effective means of efficiently recognizing patterns in data and can be applied for accurate predictions of a performance index based on investigating inputs; it can also be used to optimize leaf area index based on measurements of leaf length and leaf width.