Approach to Determining the Boundaries of the Search Range for the Number of Trees in the Random Forest Algorithm
Liliya Anatolievna Demidova, Maria Ivkina · 2020
The problem of determining the search ranges for the optimal values of the trees' number for the random forest (RF) in order to reduce the time spent on its development has been considered. The aim of the work is to obtain the formulas for determining the search range for the number of trees. The formulas have been obtained based on the results of the experimental studies on the development of the RF classifiers based on various datasets from the machine learning data repositories. The results of experimental studies on the development of the RF classifiers using the training and test sets based on the analyzed datasets have been presented. The formulas for the graphical dependencies have been obtained in the general form for assessing the classification quality at the test set and the development time.