Estimating Aqueous Solubility Directly From Molecular Structure Using Machine Learning Approach

Aqueous solubility (AS) is defined as balancing the pure compound in its gaseous, liquid, or solid phase at a specific temperature and pressure. An organic molecule may or may not like being bounded by water molecules. It is imperative to assess its environmental behavior. To understand how an organic compound dissolves in water, one must know the molecular forces and interactions. The solubility of molecules is an essential physical-chemical property for drug discovery as performed by pharmaceutical companies and academia. Aqueous solubility (AS) is a critical physical property that has been used by a medicinal as well as an agrochemical chemist. Solubility distributes biologically active compounds, thus moving the possible productivity. The same steadiness experiment to determine solubility is very time-consuming. To be able to calculate solubility when a physical sample is not present is helpful. This paper has built three linear models for predicting the AS directly from the molecular structure. This research work has compared three models based on their predictability accuracy.