> ## Documentation Index
> Fetch the complete documentation index at: https://docs.fincept.in/llms.txt
> Use this file to discover all available pages before exploring further.

# Feature engineering

> The feature matrix the models train on (lags, rolling/seasonal/expanding/EWM statistics, calendar, static and dynamic exogenous columns, per-step targets), lags and rolling statistics of exogenous series, Fourier seasonality, trend and calendar terms with their future values, and the future rows a forecast needs.

Toolset `forecast_features`: 4 tools.

| Tool | What it does | Notes |
| - | - | - |
| `forecast_feature_matrix` | Builds the exact training matrix the forecasting tools use: target (or one target per step ahead under the direct strategy), lags, rolling/seasonal/expanding/EWM statistics, calendar features and exogenous columns, after the target transforms, with incomplete leading rows dropped. | |
| `forecast_future_frame` | The (series, period) rows of the next h periods after each series' end: the rows future\_exog must supply for dynamic exogenous columns. | |
| `forecast_exog_lags` | Lags and per-series rolling/seasonal/expanding/EWM statistics of exogenous series (prices, rates, indicators), so a forecast can use their past values without needing their future ones. | |
| `forecast_calendar_features` | Deterministic regressors for history and horizon: Fourier sine/cosine seasonality terms, a trend counter, calendar attributes of the dates, and known-in-advance columns shifted into history. | |

Inputs, limits and outputs are described in [Fincept Forecast](/guides/fincept-forecast). Full schemas: `fincept_describe_tool`.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.