Enhanced Binary Tree Growth Algorithm with Linear Discriminant Analysis for Leukemia Dataset Classification

In this study, a novel meta-heuristic method named the Quasi-Opposition-Based Learning Tree Growth Algorithm (QOBL-TGA) is introduced to overcome the issues of slow convergence and local optima commonly associated with the original Tree Growth Algorithm (TGA). By combining the quasi-opposition (QOBL) and opposition-based learning (OBL) techniques, the QOBL-TGA approach enhances the global search capabilities of TGA. To assess its effectiveness, the QOBL-TGA is applied to the leukemia dataset using the Linear Discriminant Analysis (LDA) classifier. The results demonstrate that the QOBL-TGA approach surpasses the original TGA and other existing algorithms, exhibiting faster convergence and superior optimization performance. In summary, the QOBL-TGA method presents a valuable solution for addressing optimization challenges in high-dimensional problems.