Akademik Veri Yönetim Sistemi
IAENG International Journal of Applied Mathematics, cilt.56, sa.2, ss.817-828, 2026 (Scopus)
Molecular topology plays very important role in predicting the pharmacological properties of drugs. Many topological indices, computed on molecular graphs will be used for understanding the structural and biological correlations of chemical compounds. In this study, we compute degree-based topological indices including the Randic type SDI index, Exponential Fraction index, Lodeg index and Hadi index for some migraine drugs. Computational techniques were applied to survey the molecular structures and curvilinear regression analysis was performed to measure their predictive ability in modeling of drug action and affects. The results exhibits the significant correlations between the computed indices and key chemical-physical properties. We have used machine learning to improve QSPR models for migraine drugs. Traditional methods often struggle to capture complex molecular interactions, whereas machine learning can efficiently handle such nonlinear relationships. The accuracy and reliability of ML-based models are estimated by comparing their performance with traditional QSPR approaches.