Factor Preparation

Factor preparation turns raw factor observations and prices into clean long-form factor data with forward returns and quantile labels.

import ferric_alpha as fa

factor_data = fa.get_clean_factor_and_forward_returns(
    factor=factor,
    prices=prices,
    quantiles=5,
    periods=[1, 5, 10],
    max_loss=0.35,
)

Inputs

• factor: factor values indexed by date and asset-like keys.
• prices: price observations used to compute forward returns.
• quantiles: integer count or explicit quantile edges.
• periods: observed-session forward windows.

Output Contract

The output is a Polars DataFrame with date, asset, factor, factor_quantile, and one or more forward_return_* columns.

Ferric Alpha keeps missing terminal prices and invalid price windows explicit as nulls instead of silently imputing forward returns.