Details

The Feature Engineering subsystem in AlphaPy is responsible for transforming raw or ingested data into a structured and enriched set of features suitable for machine learning models. This involves creating various feature types (numerical, statistical, interaction) and applying necessary data transformations.

alphapy.features.create_features

This component acts as the primary orchestrator for the entire feature generation process. It is the main entry point for transforming raw input data into a comprehensive set of machine learning features. Its responsibilities include managing the application of different feature types (e.g., numerical, statistical, interaction) and handling data preparation steps like missing value imputation.

alphapy.features.apply_transforms

This component is responsible for managing the sequential application of individual data transformation functions. It functions as a dispatcher, iterating through a list of specified transformations and applying each function from the alphapy.transforms module to the data. This ensures a modular and extensible transformation pipeline.

alphapy.transforms

This component serves as a library of atomic, reusable data transformation functions. Each function within this module performs a specific mathematical or statistical operation on data series (e.g., log transformation, standardization, differencing). These functions are the fundamental building blocks for more complex feature engineering operations.