Особливості формування спеціалізованого набору даних для прогнозування хімічних властивостей фторвмісних сполук

О. Ф. Шевчук, П. П. Павлюк, О. Ф. Шевчук et al. · Repository of Vinnytsia National Technical University (Vinnytsia National Technical University) · 2026

The paper presents the features and results of forming a specialized dataset for predicting the physicochemical properties of saturated fluorine-containing compounds. The proposed dataset includes experimental pKa and LogP values for 183 compounds and covers several structural classes relevant to medicinal chemistry. Data preprocessing was performed, including outlier removal, stratified splitting, and molecular similarity control. Preliminary analysis revealed nonlinear relationships between the number and position of fluorine atoms and the properties of the compounds, which justifies the use of machine learning methods for further modeling.

Cited by 0 works. Open-access PDF

Authors