Regression-Based QSPR Modelling of Functionalized Aromatic Compounds with Degree-Based Topological Indices

Abstract

Degree-based topological indices (DBTI) are important mathematical tools to represent molecular structure and support quantitative structure-property relationship (QSPR) modelling. In this work, we explore the predictive capacity of DBTI for a range of functionalized aromatic compounds, which are of great importance in pharmaceutical chemistry. By combining correlation analysis with regression modelling, we establish reliable relationships between molecular topology and physiochemical behaviour. In addition to univariate regression approaches (linear, quadratic, cubic, exponential and logarithmic), multiple linear regression (MLR) was adopted to capture the combined effects of the indices. The constructed multivariate models illustrated superior predictive accuracy and robustness, evidencing the advantage of integrating multiple descriptors simultaneously. Statistical validation confirms the robustness of the developed models, demonstrating their potential as efficient computational alternatives to experimental property determination. The findings underline the significance of DBTI in predictive modelling and cheminformatics, offering valuable foresights for rational drug design and pharmacological research.