Molecules Dataset Collection: SMILES-Based Property Inference Validation

GLambard published this collection of molecular datasets in 2018 to facilitate the validation of property inference models. The data utilizes SMILES strings and RDKit-compatible formats to represent chemical structures for machine learning classification tasks.

Use Cases

• Predicting chemical properties from SMILES strings using RDKit descriptors
• Validating molecular property inference models against MoleculeNet benchmarks
• Binary or multi-class classification of molecular activity based on structural features

Strengths

• Uses SMILES notation for standardized chemical structure representation
• Compatible with RDKit and MoleculeNet standards for cheminformatics
• Released under the permissive MIT license for open research

Limitations

• Unknown total record count and specific property labels
• Dataset has not been updated since 2018, potentially missing newer molecular benchmarks
• Lack of detailed documentation regarding the specific source datasets included in the collection

Provenance

Source

GLambard via GitHub

Time Range

Up to 2018

Freshness

Last updated June 2018; no recent updates or maintenance indicated.