Some novel neighborhood degree sum-based versus degree-based topological indices in QSPR analysis of alkanes from n-butane to nonanes

The topological indices are reported to be very useful in various QSPR studies as these indices can be utilized as a tool to predict the physicochemical properties of a diverse set of chemical compounds. The objective of this communication is twofold. First, to introduce some neighborhood degree sum-based topological indices and perform a comparative QSPR analysis of the indices with their corresponding degree-based topological indices to analyze the usefulness of the new indices. Second, to investigate isomer discrimination ability of the newly defined indices on the basis of degeneracy test. Neighborhood degree sum-based topological indices not only consider the degree of each atom in the molecule but also the degrees of its neighboring atoms. Moreover, these indices are expected to provide information about how connected and dense the neighborhoods of atoms are within the molecule. In this study, we observe that the proposed neighborhood degree sum-based topological indices exhibit better correlation with some physicochemical properties of Octane Isomers and 66 alkanes as compared to their corresponding degree-based topological indices. Also, the supremacy in terms of sensitivity of the neighbourhood degree sum-based topological indices as compared to the classical degree-based ones is established. The manuscript provides an efficient and effective model for studying the structure-property relationships of some alkane molecules. Some groups of alkanes ranging from n-butane to nonane are considered for the comparative QSPR analysis. Out of the variety of topological indices that exist in the chemical graph theory literature, we consider the indices that are defined in terms of (i) Degree of vertices, and (ii) Sum of the neighborhood degrees of vertices. Some physicochemical properties (such as molecular volume, critical temperature, the heat of vaporization etc.) of the alkanes are considered to conduct the regression analysis of the TIs. To investigate the structure-property relationships of the alkanes, we perform linear regression analysis of the topological indices with some physicochemical properties of the alkane groups. The linear models of the regression analysis are validated to assess the performance by finding the values of R2 (coefficient of determination), F-test and MSE (mean squared error). The outcomes of the comparative regression analysis of the degree-based versus neighborhood degree sum-based indices with respect to the properties of the alkanes are discussed and interpretations are made. To prevent the overfitting of the models, cross-validation is carried out. Also, a degeneracy test of the TIs is performed to check the isomer discrimination ability. The prediction abilities of the degree-based indices and the neighborhood degree sum-based indices in connection with the properties of the alkanes are plotted with the help of a bar graph. image