Comparing chemical fingerprints for ecotoxicology
Résumé
We present a comparison between 14 chemical ngerprints using 1D/2D features and PMCSFG, a
data mining method that induces features based on maximum common subgraphs. We provide an experimental
evaluation and discuss the usefulness of the di erent methods on ecotoxicology data. The features generated by
data mining yield a similar performance for predicting toxicity, while they are more interpretable by chemists
Domaines
Intelligence artificielle [cs.AI]| Origine | Fichiers produits par l'(les) auteur(s) |
|---|
Loading...