feature processing

Pre Process

• steps as handling missing data

• outlier rejection

Feature Engineering

• Feature Transform / Feature Extraction

  • Applying operators to generate additional features out of the given features, the operations can be:

    • Polynomial,

    • Statistical,

    • Normalization,

    • Change of Coordinates

• Dimensionality Reduction - reduce the number of features to maximize the inner structure of the data.

• Feature Selection - only a sub set of the features are used, selected by statistical tests or closed loop evaluation.