Module ml 

Lightweight ML property proxies.

Descriptor-based property prediction using pre-fitted linear models. These provide fast estimates when full quantum-chemical calculations are too expensive.

Re-exports

pub use descriptors::compute_3d_descriptors;

pub use descriptors::compute_descriptors;

pub use descriptors::Descriptors3D;

pub use descriptors::MolecularDescriptors;

pub use ensemble::compute_tpsa;

pub use ensemble::predict_ensemble;

pub use ensemble::EnsembleResult;

pub use ensemble::VeberResult;

pub use models::predict_properties;

pub use models::MlPropertyResult;

pub use models::PredictionUncertainty;

pub use pharmacophore::compute_pharmacophore_fingerprint;

pub use pharmacophore::detect_features;

pub use pharmacophore::pharmacophore_tanimoto;

pub use pharmacophore::PharmFeature;

pub use pharmacophore::PharmFeatureType;

pub use pharmacophore::PharmacophoreFingerprint;

Modules

advanced_models

Advanced ML models: Random Forest, Gradient Boosting, and cross-validation.

descriptors

Molecular descriptors for ML property prediction.

ensemble

Ensemble ML models with non-linear predictions and uncertainty estimates.

getaway

GETAWAY (GEometry, Topology, and Atom-Weights AssemblY) descriptors.

models

Pre-fitted linear ML models for fast property estimation.

pharmacophore

Pharmacophore fingerprints: encode molecular pharmacophoric features.

rdf_descriptors

RDF (Radial Distribution Function) molecular descriptors.

whim

WHIM (Weighted Holistic Invariant Molecular) descriptors.