" Predictive Modeling of Anatomical Therapeutic Chemical (ATC) Classification Using Machine Learning on Molecular Data_ DATA"

"Dataset TitleSMILES-Based Molecular Dataset for Anatomical Therapeutic Chemical (ATC) Drug ClassificationKeywordsAnatomical Therapeutic Chemical (ATC); SMILES; Molecular Graphs; Drug Classification; Graph Convolutional Networks; Cheminformatics; Machine Learning; One-Hot Encoding; Pharmacological Data; Drug DiscoveryMethod of Data CollectionThe dataset was constructed by collecting drug compounds with assigned Anatomical Therapeutic Chemical (ATC) codes from publicly available biomedical resources, including the KEGG drug database. Molecular structures were represented using SMILES notation and curated using cheminformatics tools such as RDKit and OpenBabel. Compounds lacking valid structural information or containing ambiguous isomeric representations were excluded to ensure data consistency. Each compound was mapped to one of 14 top-level ATC classes (A\u2013V), and class labels were encoded using a one-hot encoding scheme. To support supervised machine learning experiments, the dataset was validated for structural correctness and organized in tabular format, where each row represents a unique molecule and each column represents either molecular input or ATC class membership."