Machine Learning for the Estimation of Diameter Increment in Mixed and Uneven-Aged Forests

Estimating the diameter increment of forests is one most important relationships in forest management and planning. The aim this study was to provide insight into application two machine learning methods, i.e., multilayer perceptron artificial neural network (MLP) adaptive neuro-fuzzy inference system (ANFIS), for developing models Hyrcanian forests. For purpose, diameters at breast height (DBH) seven tree species were recorded during inventory periods. trees divided four broad groups, including beech (Fagus orientalis), chestnut-leaved oak (Quercus castaneifolia), hornbeam (Carpinus betulus), other species. each group, a separate model developed. k-fold strategy used evaluate these models. Pearson correlation coefficient (r), determination (R2), root mean square error (RMSE), Akaike information criterion (AIC), Bayesian (BIC) utilized RMSE R2 MLP ANFIS estimated groups ((1.61 0.23) (1.57 0.26)), ((1.42 0.13) (1.49 0.10)), ((1.55 0.28) (1.47 0.39)), ((1.44 0.32) (1.5 0.24)), respectively. Despite low determination, test both techniques significant 0.01 level all groups. In study, we also determined optimal parameters such as number nodes or multiple hidden layers type membership functions modeling Comparison results showed that oak, technique performed better have deep relationship with nature