Estimativa da altura total de árvores utilizando diferentes técnicas de inteligências artificial

Abstract

The objective of this study was to evaluate the performance of four artificial intelligence techniques to estimate the total height of trees, using as independent variables Age, Area per plant and diameter at 1,30 m of soil (Dap). The techniques tested were Random Forest, Neuro Fuzzy, Support Vector Machine and Artificial Neural Networks, being the data processing made possible by the use of Statistica, MatLab and NeuroForest software. Two types of training were performed; for the first one the data referring to the last year of the forest inventory were not considered, and for the second one the data of all the years were used being these randomized. After the processing, the statistics Bias, RQME, correlation and average percentage error were calculated; scatter plots were generated, considering the observed values; and histogram of residues. It was verified that the processing with the randomized data presented the best statistical indices and the Support Vector Machine technique presented the best results when compared to the others.