Data preparation

This section includes topics related to feature engineering, feature selection, data validation, scaling, and more.

• Quality of a dataset
• Data validation
  • Standardizing SMILES
  • References
• Feature engineering
  • Molecular descriptors
  • Further reading
  • References
  • Contents
    • RDKit
      • Set up RDKit
      • Basic usage
      • Computing descriptors/fingerprints
      • References
    • MACCS fingerprints
      • Computing MACCS keys
      • References
    • Morgan ECFP fingerprints
      • Computing Morgan ECFPx
      • References
    • Mordred
    • Measure molecular similarity
      • Computing similarity from ECFP4
      • Another, more complete, approach
      • References
• Feature selection
  • Filter methods
    • Collinearity (correlation metric)
  • Wrapper methods
    • Recursive feature elimination
      • Drawbacks
    • Boruta
      • Advantages
      • Drawbacks
    • References
  • Embedded methods
    • Selection by model importance
      • Advantages
      • Drawbacks
  • Dimensionality reduction methods
  • Hybrid methods
• Train, validation, and test sets
  • Training set
  • Validation set
  • Testing set
  • Conducting the split in sklearn
• Scaling
  • Is feature scaling always necessary?
  • Normalization
  • Standardization
  • Scaling data before or after train/test split