Transforms

Feature Transforms

Featurizer provides a set of built-in operators for common transformations on time-series data called Transforms

Some examples are:

• diff - transforms time-series into a series of differences between sequential events
• min_max_scaling - scales values between 0 and 1 based on min and max value in a window
• ewma - exponential smoothing

Transforms are similar to regular FeatureDefinitions when coming to definition (they subclass FeatureDefinition), however differ in how they are used in FeaturizerConfig: unlike regular FeatureDefinition which do not need direct specification of dependant features (they are constructed automatically), Transforms require specific definition of dependencies:

feature_configs:
  - feature_definition: transforms.diff
    name: diff_mid_price
    deps:
      - mid_price
    params:
        ...
  - feature_definition: price.mid_price
    name: mid_price
    params:
        ...

Creating new transforms

WIP