Akademik Veri Yönetim Sistemi
V. INTERNATIONAL APPLIED STATISTICS CONGRESS (UYIK - 2024), İstanbul, Türkiye, 21 - 23 Mayıs 2024, ss.72
Gene selection is a feature selection process that is widely used in machine learning, data mining and pattern analysis. Gene selection problem has a large number of genes which include relevant, redundant and noise. Developments in bioinformatics show that reducing the number of genes is an important factor for more effective and faster diagnosis of diseases. In this problem, computational complexity is reduced by selecting small number of genes. However, it is necessary to choose the relevant genes in order to maintain high level of accuracy. In this purpose, researchers have proposed several feature selection (FS) methods to select the relevant genes. In this study, the binary tuna swarm optimization algorithm (BTSO) is proposed and applied to gene selection problem. Thus, BTSO select the most relevant genes for a disease from microarray data. The performance of the proposed method is demonstrated by applying various classifiers to the selected gene subsets and different evaluation criteria such as accuracy, recall, precision, F1-score and confusion matrix are used. The experimental results demonstrate that the proposed method can reduce the dimensionality of the microarray dataset, identifying the most informative gene subset, and improving classification accuracy.