Artificial Intelligence for Drug-Likeness Prediction: An Intelligent Framework for Chemical Compound Screening
TL;DR
An AI-based framework for predicting drug-likeness using molecular descriptors derived from chemical structures is proposed that integrates data preprocessing, feature selection, machine learning model development, and performance evaluation using metrics such as accuracy, precision, recall, F1-score, and AUROC.
Abstract
Drug discovery requires screening large numbers of chemical compounds to identify promising drug candidates. This study proposes an AI-based framework for predicting drug-likeness using molecular descriptors derived from chemical structures. The framework integrates data preprocessing, feature selection, machine learning model development, and performance evaluation using metrics such as accuracy, precision, recall, F1-score, and AUROC. An explainability component is also included to identify important molecular properties influencing predictions. The proposed approach can support early-stage compound screening by prioritizing potentially suitable drug candidates and reducing the need for costly experimental analysis
