Journal of Advances in Mathematics and Computer Science

Misbalance Adriatic topological indices quantify differences between the degrees of adjacent vertices and may serve as molecular descriptors in quantitative structure-property relationship (QSPR) modelling. This study evaluated the predictive performance of four such indices—the misbalance deg index (MD), misbalance rodeg index (MRD), misbalance irdeg index (MIRD), and misbalance haddeg index (MHD)—for selected nonsteroidal anti-inflammatory drugs (NSAIDs). The indices were computed from hydrogen-suppressed molecular graphs and examined in relation to six physicochemical properties: boiling point, polarizability, complexity, refractivity, molecular weight, and molar volume. Pearson correlation coefficients were used to assess linear associations, and separate simple linear regression models were developed for each descriptor-property combination. All four indices showed positive correlations with the investigated properties. Polarizability and refractivity generally exhibited the strongest associations, whereas molar volume showed weak correlations across all descriptors. MD provided the strongest overall regression performance and the highest coefficients of determination for boiling point, complexity, and molecular weight. MRD performed best for polarizability and refractivity, while MHD produced the highest coefficient of determination for molar volume, although this model remained weak. MIRD and MHD yielded broadly comparable results for the remaining properties. The findings indicate that the investigated indices may support preliminary QSPR estimation within the analysed NSAID dataset, while their predictive utility requires further assessment using larger datasets and additional modelling approaches.