<![CDATA[Feature selection is one of the most crucial concepts of learning when constructing a machine learning algorithm. This paper proposes a new technique of using a blend of Secretary Bird Optimization (SBO) and Decision Tree for feature selection. SBO, with the consideration of hunting strategy of called secretary bird, can successfully search the space and find feature subsets. The first proposed framework involves identifying the relevant features by applying SBO then secondly deciding on a ranking of the features by using a decision tree model classifier. Evaluation is based on the most famous Iris dataset with the application of the 5-fold cross-validation to increase the reliability of the results. It is shown that the SBO-based approach succeeds in both objectives, and the average classification accuracy is equal to 0.9550 ± 0.0316, while the baseline selection methods have higher values of loss. This result has unveiled a theoretical and practical potential for future works that seek to combine metaheuristic optimization with decision trees and interpretability of selected features and machine learning models.]]>
Copyrights © 2025
