Bonding alkane attributes with topological indices: a statistical intervention

Abstract

This research delves into a comprehensive investigation of a specific group of alkanes, with the primary objective of establishing meaningful associations between their inherent physical attributes and a set of graphical parameters ranging from topological indices to their associated graphical entropies. Though there are countless techniques to relate molecular structure and physical properties of chemical compounds, the pursuit has often been pricey and arduous. With a view to curb the expense and intense labour, this work focuses on attaining some of these properties using graphical interpretations and statistical interventions. Driven by this objective, we compute Sombor index, a recently developed topological tool and some of its variants for all structural isomers of alkanes with carbon range four to nine. Furthermore, our study employs multiple regression analysis to explore their connections with some physical attributes of alkanes while concurrently addressing the challenge posed by multicollinearity. We have adopted the technique of robust regression to scale down the impact of outliers in the dataset, while generating regression models. Further, the established results are justified with a 10-fold cross validation process and a comparison with the results obtained from different topological indices. The results of this research contribute a valuable insight to the fields of chemical informatics and structural analysis, offering an enhanced comprehension of the interplay between molecular structure and physical attributes within alkanes.