Utilizing decision tree models for estimating wood extraction methods

Wood extraction (WE) in forestry is a difficult, dangerous and costly process necessary for the sustainable management of forests. Experts or expert systems are needed to decide on the method in the mountainous forest areas of our country where there are variable terrain conditions. These processes can become easier with artificial intelligence (AI). The aim of this study is to determine the most suitable extraction method by considering the variable conditions of forest areas, and to develop an AI system to estimate the duration of the work and production costs according to the determined method. The daily production (m3/day) data obtained from the studies conducted in Artvin, one of the richest directorates of our country in terms of mechanization, and the data set formed with expert opinions were used as training and test data for the decision tree (DT) in RapidMiner Program. skidding distance, slope of the terrain, direction of transportation/skidding, terrain obstacle, weed, soil condition and the amount of product to be extracted were taken into consideration in the evaluation of the WE. With the model, which has 94,60% success in predicting the WE method, controls were carried out on 6 similar land pairs with and without planning in the production studies of 2023. As a result of the studies, the success of the model in predicting the WE method was determined as 100%.