Studying diameter distribution of natural secondary stand based on artificial neural network

Abstract

Taking the investigated data from 40 samples plots of natural secondary oak stand in Baotianman natural reserve for the research object, BP-ANN model was created by using relative diameter of tree as the input variable, and accumulated frequency of tree number as output variable. Through training and optimal seeking by the software of MATLAB, the idea network model was created. In the performance analysis of 7 models of main tree species, the fitting accuracy is 96.92% to 100%; in the test analysis of the created model by not used data, the test accuracy is 97.95%; and in the X2 test, the fitting effect is remarkable. The results indicated that ANN method is a more effective way to model diameter distribution of natural secondary oak stand.