Comparative entropy analysis of 2D transition metal tetrahydroxyquinones via machine learning approaches

This paper focuses on the study of the framework of transition metal tetrahydroxyquinone (TM-THQ) which is a chemical belonging to the metal-organic framework (MOF) family. These materials are becoming relevant due to their stability and high surface area that makes them prospective in terms of carbon-dioxide conversion and other uses. To explore the structure of TM-THQ, we apply molecular graph theory, in which atoms and bonds are represented as mathematical graphs. We determine some topological indices, the Zagreb indices, Randić index, the atom-bond connectivity index, the geometric-arithmetic index, and the sum-connectivity index to explain the connection between atoms and predict the stability and reactivity of the material used. Moreover, Shannon's entropy model is applied to estimate the entropy of the indices. To this end, we develop and compare three machine learning based regression models: logarithmic, random forest, and XGBoost. All computational work, data analysis, and machine learning modeling were performed using Python. The findings reveal a strong predictive relationship, indicating that the entropy of TM-THQ can be accurately forecasted from its topological descriptors, with the ensemble methods providing superior predictive performance.