Document Type : Research Paper
Authors
1
M.Sc. Graduate, Dept. of Forestry and Forest Economics, Faculty of Natural Resources, University of Tehran, Karaj, Iran.
2
Professor, Dept. of Forestry and Forest Economics, Faculty of Natural Resources, University of Tehran, Karaj, Iran.
3
Professor, Dep. of Agricultural Engineering and Technology, Faculty of Agriculture Machines Engineering and Natural Resources, University of Tehran, Karaj, Iran.
4
Assistant Prof, Research Institute of Forests and Rangelands, Agricultural Research, Education and Extension Organization (AREEO), Tehran, Iran.
Abstract
The use of statistical experimental models is common practical methods among forest resource managers. Regression analysis is also a statistical method that can be used to estimate the volume. This method has limitations and requires assumptions such as normality, homogeneity of variance and non-linear relationship. The use of new techniques such as artificial neural networks can deal with these limitations. This study aims at comparing the performance of Artificial Neural Networks (ANN) and regression analysis to estimate the total volume of logs. For this purpose, 367 trees out of marked trees in research and educational forest of kheyroud were selected and DBH, diameter at stump height, stump height, total tree height, species, tree situation, minimum median diameter and topographic factors such as aspect and elevation were measured with high accuracy. Multilayer perceptron (MLP) and multivariate regression were developed to estimate the total volume of logging trees. The results indicated that the Neural Network was more accurate about 40% in estimating the total volume of logging trees than the regression method. Comparing evaluation criteria showed RMSE value 1.411 for ANN modeling and 3.49 for regression analysis. The difference between estimated and actual total volume was 6.5% to regression analysis and 1.7% to Neural Network. According to the results, the amount of difference was less for ANN model than regression model.
Keywords
)