Efficiency of gene expression programming in diameter-height modeling of Iranian oak (Quercus brantii Lindl).

measuring the height of all forest trees is a time-consuming and expensive operation, hence the use of diameter and height models to estimate the height of trees has been developed. The aim of this research is to investigate the efficiency of gene expression programming in diameter-height modeling of Iranian oak species (Quercus brantii Lindl) in high forests of Middle Zagros.

In order to carry out this research, by conducting numerous forest walk and getting to know the forests of the region, a stand with an area of approximately 5 hectares with a high forests vegetation structure was selected in the protected area of Sefid Koh Lorestan became. In this stand, the characteristics of DBH and total height of all Iranian oak trees whose DBH was more than 12.5 cm were counted 100%. A total of 642 trees were measured. In this research, 80% of the data was used for modeling and 20% for validation. The gene expression model with 3 genes and 100 chromosomes was implemented to investigate the relationship between height as a dependent variable and diameter as an independent variable. In order to evaluate the performance of the final model, RMSE, MAE and R2 criteria were used.

The model extracted from GEP justified 87% of the tree height based on the R2 value. The results of tree height-diameter modeling showed that the final model obtained has coefficient of explanation (R2), root mean square error (RMSE) and mean absolute value of error (MAE) 0.87, 1.3 and 0.97 respectively. is Also, the results of the criteria used to validate the obtained model showed that the extracted model has the coefficient of explanation (R2), root mean square error (RMSE) and mean absolute value of error (MAE), respectively 0.82 and 40. 1 and 1.06 could predict the height of trees.

Overall, the results of this research showed that the model extracted from gene expression programming according to the R2, RMSE and MAE performance evaluation criteria has the ability to estimate the height of Iranian oak high forests in the middle Zagros vegetation zone. Therefore, this model can be used in the forest areas of the middle Zagros vegetation zone, which have the same structure and habitat conditions as the studied area. It should be noted that the present research only predicted the height of trees based on the DBH (independent variable), so it is suggested that in future researches, the generalized models of height and diameter, in which the variability of the habitat and stand Considering the stand variables other than tree diameter (basal area of stand, stand age, dominant height, dominant diameter, site index, etc.) are considered to be used. Also, in order to make a more accurate judgment about the performance of this program, it is better to compare and measure it with other estimation algorithms such as non-linear regressions, support vector machine, artificial neural network, etc. in future studies.