Forecasting relative tree growth based on PPR artificial neural network model

Relative growth is defined as the description of the growth relationship between biological and biological part (organ). The relative growth function is a function on variable of growth and non-time variables. In the paper, a system relative growth model which taken Betula costata as an example was constructed based on the projection pursuit regression (PPR) artificial neural network. The system model was constructed in order to predict the tree growth. The results show that the artificial neural network model were available for prediction of the tree growth; the predication average relative error of diameter growth was 0.04, the predication average relative error of height was 0.06,the predication average relative error of volume was 0.12. The PPR model was available for tree growth predication which can be applied to predict the tree growth.